00:01:16 *** dmiles_afk (i=dmiles@c-67-170-65-58.hsd1.wa.comcast.net) has joined #swhack
00:01:16 <Monty> it's dmiles_afk!
00:05:54 *** lordi has quit (Remote closed the connection)
00:08:04 <Morbus> phenny, tell sbp http://www.dragonrest.net/fortfiles/forthome.html
00:08:10 <phenny> Morbus: I'll pass that on when sbp is around.
00:08:51 *** dmiles_afk has quit (Read error: 104 (Connection reset by peer))
00:09:25 *** dmiles_afk (i=dmiles@c-67-170-65-58.hsd1.wa.comcast.net) has joined #swhack
00:12:59 *** BigJibby (n=matt@142.46.8.22) has joined #swhack
00:24:50 <zachb> .title www.the-end.com
00:24:52 <phenny> zachb: The Prophesied End-Time Revealed: Ronald Weinland's Books
00:24:52 <Monty> angry something
00:35:57 <bancus> .u element
00:35:58 <phenny> U+2208 ELEMENT OF (∈)
00:36:35 <bancus> .u subset
00:36:36 <phenny> U+2282 SUBSET OF (⊂)
00:36:42 <bancus> .u proper subset
00:36:44 <phenny> bancus: Sorry, no results for 'proper subset'.
00:38:52 <bancus> .u open subset
00:38:54 <phenny> bancus: Sorry, no results for 'open subset'.
01:24:00 *** shoulson_ is now known as clsn
01:24:02 <_bjoern> .gs All I need is this *
01:24:05 <phenny> All I need is this *: guitar (4), song (3), a monocle (3), track (2), satisfaction (2), information (2), chair (2)
01:24:14 <_bjoern> .gs All I need is that *
01:24:15 <phenny> All I need is that *: q9450 (3), then (2), fiddle (2), cts coupe (2), cheap ethernet (2), chance (2), canon (2)
01:24:17 <_bjoern> .gs All I need is your *
01:24:18 <phenny> All I need is your *: love (12), address (4), word (3), touch (3), permission (3), loving (3), heart (3), bank (3), vote (2), loveeeee (2), body (2), approval (2)
01:24:19 <_bjoern> .gs All I need is his *
01:24:20 <phenny> All I need is his *: love (12), name (3), attention (3), word (2), w (2), trustworlthy (2), arch-enemy kampfer (2)
01:24:22 <_bjoern> .gs All I need is her *
01:24:23 <phenny> All I need is her *: smile (3), name (3), tenderness (2), sweet (2), speacil (2), signature (2), shape (2), presence (2), in (2)
01:24:25 <_bjoern> .gs All I need is its *
01:24:26 <phenny> All I need is its *: way, uncanny resemblance, soul, radius, info in my dex, importance)
01:24:57 <_bjoern> .g "arch-enemy kampfer"
01:24:58 <phenny> _bjoern: http://ngeekhiong.blogspot.com/2006/05/plamo-review-5-sd-gundam-nt-1.html
01:25:09 <_bjoern> .wik Gundam
01:25:10 <phenny> "Mobile Suit Gundam was developed principally by renowned animator, Yoshiyuki Tomino, along with a changing group of Sunrise creators who went under the collective pseudonym 'Hajime Yatate'." - http://en.wikipedia.org/wiki/Gundam
01:29:46 *** Xanthor[aw] has quit (Read error: 110 (Connection timed out))
01:32:46 *** Xanthor[aw] (n=Xanthor@ALyon-257-1-159-84.w81-251.abo.wanadoo.fr) has joined #swhack
01:45:27 <_bjoern> .gcs [hard porno] [soft porno]
01:45:29 <phenny> hard porno (1,070,000), soft porno (520,000)
01:45:37 <_bjoern> .gcs [hard porn] [soft porn]
01:45:37 <phenny> soft porn (3,600,000), hard porn (3,080,000)
01:47:28 <_bjoern> need photos.cx replacement, quick!
01:48:32 *** skwerl is now known as melbel
01:52:05 *** tr_ has quit (Read error: 104 (Connection reset by peer))
01:53:35 <chandler> _bjoern: try tinypic.com
02:00:38 <_bjoern> 2late
02:02:03 <bancus> .gcs [hardcore porn] [softcore porn]
02:02:04 <phenny> softcore porn (1,930,000), hardcore porn (469,000)
02:23:51 *** jeffarch has quit (Read error: 110 (Connection timed out))
02:25:53 <_bjoern> MONTY DONT MISS ME WHILE IM GONE
02:25:58 <Monty> "<meta http-equiv='X-UA-Compatible' content='YOUR MOM'>" - http://en.wikipedia.org/wiki/Voltage
02:25:59 <zachb> That's odd
02:26:11 <_bjoern> You don't say!
02:26:12 <zachb> _bjoern: you're going?!
02:26:25 <_bjoern> it's 0430h here.
02:26:26 <zachb> It was @ the softcore porn > hardcore porn
02:26:28 <zachb> oh
02:26:34 <zachb> I thought you were leaving me FOREVER
02:26:44 <_bjoern> I would never do that!
02:26:48 <_bjoern> Having never been with you!
02:26:53 <zachb> I'll never leave you either!
02:27:08 <_bjoern> Thank you, most disturbing!
02:27:18 <zachb> I'm never gonna give you up!
02:27:39 *** nsh has quit (Read error: 104 (Connection reset by peer))
02:27:50 <_bjoern> Not even under the threat of nuclear annihilation?
02:27:56 *** nsh (n=nsh@87-94-146-186.tampere.customers.dnainternet.fi) has joined #swhack
02:28:24 <zachb> Nope, I'd never let you down!
02:29:07 <zachb> Why no voice?
02:29:27 <_bjoern> I hate inconsistency.
02:30:32 <_bjoern> MONTY GOOD NIGHT
02:30:35 <Monty> unless age of consent to if there is 16 is the Forestâ?)
02:30:41 <_bjoern> MONTY BITE ZACHB THE MOMENT IM GONE
02:30:44 <Monty> lol http://cgi.4chan.org/gif/src/1212247572293.gif
02:30:52 *** kpreid has quit ()
02:31:08 <_bjoern> I know better than to click on 4chan links this hour
02:31:19 *** _bjoern has quit ("Quit")
02:31:31 <bancus> It's probably 404d by now if Monty is repeating it.
02:31:33 <Monty> if (Win2k == Berk) { flatmates hates sweet GCSE;}
02:31:40 *** jewel has quit (Read error: 113 (No route to host))
02:32:00 *** kpreid (n=kpreid@cpe-69-202-162-8.twcny.res.rr.com) has joined #swhack
02:34:47 *** jeffarch (n=jja@pdpc/supporter/active/jeffarch) has joined #swhack
04:22:28 *** mahound has quit (Read error: 110 (Connection timed out))
04:46:45 *** tr_ (n=c@c-98-216-10-223.hsd1.ma.comcast.net) has joined #swhack
04:48:32 <zachb> _bjoern! You're back! I was never going to run around, or even desert you, but you came back!
05:39:50 *** tr_ has quit (Remote closed the connection)
05:41:23 *** tr_ (n=c@c-98-216-15-55.hsd1.ma.comcast.net) has joined #swhack
05:58:49 <Mike_L> hmm
06:01:09 * Mike_L is thinking about backup services that provide strong data retension guarantees
06:02:13 <Mike_L> specifically, I want to make sure that there is no automatic way to destroy the all of the old backups
06:03:15 <Mike_L> traditional backups that use removable media are cracker-safe
06:03:53 <Mike_L> but as we move to network based backup systems, I fear that safety from malicious data deletion is often overlooked
06:05:58 <Mike_L> most network backup systems based on rsync are vulnerable to malicious data deletion
06:06:40 *** tr__ (n=c@c-71-232-26-86.hsd1.ma.comcast.net) has joined #swhack
06:07:16 <xover> .wik .wik Write Once Read Many
06:07:17 <phenny> "Worm, an elongated soft-bodied invertebrate animal" - http://en.wikipedia.org/wiki/WORM
06:07:23 <xover> .wik Write Once Read Many
06:07:24 <phenny> "Write Once, Read Many (alternatively Write One, Read Multiple or Write Once, Read Mostly or WORM) refers to computer data storage systems, data storage devices, and data storage media that can be written to once, but read from multiple times." - http://en.wikipedia.org/wiki/Write_Once_Read_Many
06:08:13 <Mike_L> the best way to set up safe backups with rsync is to have a separate host pull the data, and separate the two hosts into different security zones so comprimization of one backed-up host does not lead to the compromizing of the backup server
06:08:18 <tr__> .wik wyrm
06:08:18 <phenny> "Wyrm (World of Darkness), a malefic entity in the World of Darkness role-playing games" - http://en.wikipedia.org/wiki/Wyrm
06:08:54 <xover> Now you're talking basic security practice (compartimentalization).
06:09:05 <xover> (god I hope I spelled that word wrong)
06:09:14 <Mike_L> compartmentalization
06:09:22 <tr__> cuz one theyve compromized one copy of the data, theyll want a second copy for redundancy...
06:09:42 <Mike_L> the problem with pull-rsync is that one has to configure the backups on the backup server... so configuration is split
06:10:07 <xover> Why would you consider that a configuration split?
06:10:22 <Mike_L> because you have to configure the application and the backups on separate servers
06:10:40 <xover> Hmm. You're in a single-app environment?
06:10:50 <Mike_L> I'm talking about each app
06:11:05 <Mike_L> or even each server
06:11:33 <xover> The backup system is an application; it would be odd to split its configuration between the central server and every app it takes backups of.
06:11:48 <Mike_L> it would be better to have the server push its backups and use a write-once read-many storage service
06:12:26 <Mike_L> I think it's a mistake to separate backup into a separate application
06:12:42 <xover> Why?
06:13:21 <Mike_L> because of the burden of configuring two applications
06:13:50 <Mike_L> when you set up an application, you configure all of its resources: directories, databases, memory, network ports, etc.
06:13:57 <Mike_L> I think 'backup store' should be one of those resources
06:14:25 <xover> So every application — the Frobnitz, that's the best Bamfwidget in the market, say — should /also/ be great at backup?
06:14:25 <Mike_L> and the application should be responsible for rendering a consistent image of its state to the backup store
06:14:43 <xover> Applications need to be aware of backup, yes.
06:14:51 <Mike_L> apps are already good at rendering their state to disk, to allow startup-shutdown
06:15:23 <Mike_L> but things get more complicated with databases
06:15:44 <Mike_L> and many apps store extra data that need not be backed up, such as database indexes
06:16:00 <xover> Network ports allready need to be configured on a separate system; the external firewall.
06:16:31 <Mike_L> that's because we don't yet have firewalls that allow applications to present credentials and get those ports opened automatically
06:16:46 <xover> There's a reason for that.
06:17:09 <Mike_L> why's that?
06:17:52 <xover> What happens if that server is compromised (through an SQL injection, say)?
06:18:19 <Mike_L> ok
06:18:55 <xover> That's why you use an external firewall instead (or rather, in adition to) the local OS firewall.
06:19:22 <Mike_L> the attacker can get credentials that lets it impresonate the application to the firewall
06:21:13 <Mike_L> s/impresonate/impersonate/
06:21:45 *** tr_ has quit (Read error: 110 (Connection timed out))
06:21:53 <Mike_L> it just seems to me that a backup service is a really good idea and would greatly simplify a lot of systems
06:22:31 <Mike_L> and I have a bit of experience in large systems now
06:22:59 <xover> Things like Oracle, or Microsoft Exchange, have a Backup API. The original manufacturer, third parties, or the backup software vendor, will provide a backup agent that talks to that API.
06:23:00 <Mike_L> so I can say that in my experience, an application's configuration should be deployed and tested as a single unit
06:23:36 <Mike_L> yes, and I think that's backwards. I've struggled with Exchange backup agents before.
06:24:27 <Mike_L> Microsoft exposed an API to read and write all objects in the database and then leaves it up to the agent to implement the actual backup functionality
06:24:54 <xover> A configuration should be deployed and tested as a single unit, yes; but that doesn't mean there needs to be a single huge configuration file for everything.
06:25:18 <xover> It only means you need to treat all changes, in all related systems, as one unit.
06:25:30 <Mike_L> so the bulk of the backup/restore code lives outside of Exchange, so you have sizeable bodies of code that are supposed to work together: a recipe for lots of bugs
06:26:16 <xover> That's the Microsoft philosophy; the *ix philosophy is lots of small tools that do one thing and do them well.
06:26:41 <Mike_L> I disagree
06:26:54 <Mike_L> unix is a platform, not an application
06:27:29 <Mike_L> and unix only works because it has a decent interface between applications: files and unix sockets
06:27:31 <xover> Fine, take Apple as an example; they have lots of huge apps, that do one thing very well, and still manage to play well (integrate well) with other parts of the system.
06:27:45 <Mike_L> I don't know much about Apple
06:27:54 <Mike_L> I've never owned a Mac
06:28:00 <xover> Case in point; Time Machine.
06:29:01 <Mike_L> but I think it is well accepted in software engineering that separate bits of code should not rely on the internals of other bits of code - which is the situation with Exchange and its backup agents
06:29:20 <xover> Well, yes, Exchange sucks in so many ways.
06:29:52 <xover> You can debate the relative merits of any given Backup API, any given backup agent, and any given backup system.
06:30:54 <Mike_L> doesn't Time Machine also suffer from the file-based backup systems? It needs a consistent view of database files, and will back up tons of unnecessary data like cached files and indexes
06:31:19 <Mike_L> would you like to stop debating them?
06:31:54 <xover> For being an apparently hopelessly naive system, Time Machine manages in practice to do pretty much the right thing 99.99…% of the time.
06:32:24 <xover> Partly because the OS foundation frameworks are aware of Time Machine and accomodates it, I'll grant.
06:32:47 <Mike_L> doesn't it employ filesystem-level snapshot capabilities of the OS? I would hardly call that naive
06:33:38 <tr__> i doubt it. they use HFS+ still
06:33:42 <xover> Nope. It makes use of change-tracking at the filesystem level, but uses *nix style hard links to avoid duplication.
06:33:47 <Mike_L> I'm not suggesting that *every* application should handle its own backups, only that some important applications could benefit a lot from it
06:34:00 <xover> Sure, I'll buy that.
06:34:13 <xover> cf. Oracle, and rman.
06:34:19 <Mike_L> for example, WordPress
06:34:37 <xover> Of course, trying to actually use rman for Oracle backups is a right royal pain in the ass.
06:34:48 <Mike_L> it would be great to be able to point WordPress at a backup service URL
06:34:55 <Mike_L> what's rman?
06:35:04 <xover> You can easily script it, and trying to get it to integrate into your storage system is a fucking lost cause.
06:35:15 <xover> rman is Oracle's in-app backup facility.
06:35:30 <Mike_L> then if one moved his blog to another server, you could just enter the backup service URL into the new WordPress and click 'restore'
06:35:37 <xover> (in-app here means in the database app, not the client talking to the database).
06:35:43 <Mike_L> ok
06:36:56 <xover> For one thing, when rman tries to manage your storage system it needs to have drivers for it; and Oracle being a database house and not a storage house, they have limited resources for producing drivers for storage systems.
06:37:16 <Mike_L> my minimal experience with Oracle has given me the impression that it's quite difficult to use, which makes sense because Oracle is targetting customers who will pay $25,000+ and are subsequently willing to put up with difficult-to-use software
06:38:07 <xover> Oracle is huge and very advanced, which also makes it complex and hard to learn.
06:38:19 <Mike_L> so I've been thinking about this backup service and what it will need
06:38:22 <xover> It's also far less easy to use than it could have been, had that been a priority.
06:38:55 <tr__> Mike_L: what would backup service imply more than what s3/rsync can do
06:39:36 <xover> If your case is WordPress, I would venture you're really thinking ore of portability of content and import/export than backup per se here.
06:40:08 <Mike_L> tr__: several things: avoiding duplication, preventing malicious deletion, and automatic deletion of expired data
06:40:08 <tr__> configuration portability
06:40:11 <tr__> esp wrt 3rd party modules etc
06:40:33 <Mike_L> xover: configuration is data, too
06:40:40 <tr__> custom layouts, templates, all that
06:41:06 <xover> Hmm. A relevant Time Machine feature; it's very very easy to install a new system by telling the installer to restore from a given Time Machine backup.
06:42:47 <xover> IOW, Time Machine would in that sense fill your requirements; both backup and portability.
06:43:04 <Mike_L> I think the application could avoid duplication, but this is something more suited to the backup server for two reasons: 1) checking for existing duplicates requires network traffic, 2) a generic backup service would contain one copy of the duplicate-avoidance and eliminate the code from each application
06:43:17 <tr__> portablity? between OSX machines maybe
06:43:31 <tr__> you also porably have to carry around the little time machine hd thingy (ie no cloud)
06:43:32 <xover> tr__: Within the limits of its scope, yes.
06:43:36 <Mike_L> xover: Time Machine has a lot of the same features that I'm aiming for
06:44:16 <xover> tr__: Also, Time Machine can use a networked drive for its backups.
06:44:31 <Mike_L> xover: I think that Time Machine does not prevent malicious deletion of backups
06:45:14 <xover> Correct. See “scope” above. :-)
06:45:31 <Mike_L> a backup service could easily do WORM (write-once read many) as you mentioned earlier, but this creates the problem of deleting unneeded backups
06:45:50 <xover> It also creates problems of scaling your storage system.
06:45:54 <Mike_L> how so?
06:46:18 <xover> WORM products are expensive, slow, inefficient; and hard to find these days.
06:46:21 <Mike_L> I would say it makes it *much* easier to scale, since one need not accomodate updates to existing objects
06:46:25 <tr__> Mike_L: handle backup dependency (eg incremental backup) and then you can garbage-collect time-inveral stuff
06:47:07 <Mike_L> tr__: yes, garbage collection is essentially the same problem
06:47:27 <xover> Making a software system logically write-once (i.e. in software) may make it simpler in some sense, but much harder in others.
06:47:41 <Mike_L> tr__: how can we represent dependencies between backed up objects and still keep the service generic?
06:48:09 <Mike_L> xover: please explain
06:48:49 <xover> How do you algorithmically guarantee write-once?
06:48:58 * thelsdj watches steven moffat doctor who
06:49:04 <Mike_L> xover: excuse me?
06:49:47 <Mike_L> xover: ok I'll answer that
06:50:18 <xover> If it's the software, not the hardware, that provides the “write once” property; how do you design an algorithm to actually guarantee that?
06:50:58 <Mike_L> xover: with a cluster of storage nodes and distributed datastructure mapping object-ids to blobs of data on particular servers, when a write comes in to a server you check if that object-id already exists and fail if it does
06:51:42 <Mike_L> xover: otherwise you accept the blob. If you later discover a different blob with the same object-id then you keep the one with the oldest timestamp and delete the others
06:51:59 <xover> Who manages the IDs? WHo /generates/ the IDs? What happens if theres a bug in either component?
06:52:12 <xover> If you delete a blob you've stored, the system is no longer write-once.
06:52:29 <Mike_L> xover: yes, and that's what I've been trying to get at
06:52:36 <xover> What happens if a node is off the network? Out of sync with its controller?
06:52:43 <Mike_L> there must be some way to delete the data, but you can't trust the application to do it
06:53:21 <xover> One possible way would be to keep the object. but mask it or mark it as a dup/deleted/hidden/etc.
06:54:15 <Mike_L> xover: ids are managed by a map service which runs on your cluster. The customer provides IDs. If a node is off the network then you detect the duplication and correct it later when the node comes back. You can use algorithms that are not vulnerable to synchronization problems
06:55:18 <Mike_L> xover: if you want to provide consistent responses to the client then you need a map service that has strong consistency and reliability
06:56:09 <Mike_L> xover: consistent responses mean if you respond 'OK the object with id-XXX was stored' then future requests for id-XXX will return the same data
06:56:51 <Mike_L> strong consistency and reliability are *much* easier to do for a map service than for the whole data store. This is what I meant when I said that WORM makes it easier to scale
06:57:22 <xover> What does it mean for the service to “run on the cluster”? Where is the actuall process or processes running?
06:57:49 <Mike_L> xover: it's running on several servers in the cluster
06:57:56 <xover> But not all?
06:58:14 <Mike_L> maybe all, if you're using an algorithm that can do that
06:58:15 <xover> Do all participating nodes have a complete view of the namespace?
06:58:54 <Mike_L> these are very interesting questions which I've been studying, but I'm not an expert
06:59:43 <xover> Oh, hey, don't mind me; I'm just throwing problems at you so you'll have something to think “against”. :-)
07:00:36 <Mike_L> there are lots of ways to build a map service: paxos-based services such as Google's Chubby lock service, distributed hash tables, data structures that use optimistic replication and eventual consistency. I can give you titles of a bunch of papers to read, if you want to go into this
07:00:50 *** _bjoern (n=bjoern@dslb-084-056-251-127.pools.arcor-ip.net) has joined #swhack
07:01:18 <Mike_L> but this is getting off track. Tonight I'm thinking about backup services
07:01:31 <xover> Not particularly right now, but if you collect it up into a blog post or article somewhere I'm sure I'll be most interested to read it at some point!
07:01:31 <Monty> night, anyway
07:01:33 <Mike_L> and what is the minimum functionality for a good generic backup service
07:02:21 <Mike_L> xover: I'm planning to start blogging. I installed pyblosxom on my website but haven't started using it yet
07:02:37 <xover> Start using it, fix it later! :-)
07:03:14 <Mike_L> I would use WordPress but it's a lot more complicated to backup and restore
07:03:22 <xover> (said one who's so anal he decided he had to write his own blog system before he could possibly blog anything)
07:03:46 <Mike_L> oh that reminds me...
07:04:05 * Mike_L starts his rsync pull-based backup script to backup his web server
07:04:26 <xover> Now surely that's a job for cron?
07:04:28 <Mike_L> rsync works really well on log files
07:05:33 <Mike_L> I would use cron except I'm not always running Linux and I don't leave my computer on all the time
07:05:55 <xover> Anacron; runs missed jobs when next it can.
07:06:43 *** mahound (n=mahound@unaffiliated/mahound) has joined #swhack
07:07:23 <Mike_L> cool
07:08:07 <xover> But don't use it for jobs scheduled disporpportionately more frequent than your gaps; or you'll find yourself making locking code in your cron scripts.
07:08:59 <Mike_L> eventually I want to back up to my little NSLU2 - ARM cpu & 32MB RAM & 80GB USB hdd
07:09:16 <xover> NSLU2?
07:09:22 <Mike_L> .wik NSLU2
07:09:23 <phenny> "The NSLU2 (Network Storage Link for USB 2.0 Disk Drives) is a now discontinued device that was made by Linksys." - http://en.wikipedia.org/wiki/NSLU2
07:09:33 <Mike_L> DISCONTINUED!!!?!?!?!
07:09:52 <xover> .wik Drobo
07:09:53 <phenny> "Drobo is a data storage peripheral, developed by Data Robotics, Inc., which interfaces up to four 3.5' SATA hard drives with a computer via High-Speed USB 2.0 (no FireWire or eSATA)." - http://en.wikipedia.org/wiki/Drobo
07:10:33 <xover> Haven't tested, but it looks really really nice; and it has a network attach peripheral that does, I think, about what the NSLU2 does.
07:10:44 <Mike_L> the nice thing about the NSLU2 is that you can install Linux on it and leave it on all the time. It draws very little power
07:11:10 <Mike_L> Drobo is 2x the price of nslu2
07:11:22 <xover> Pretty pricey up front, yeah.
07:12:06 <xover> I suspect over time it's cheaper, iff the redundancy it gives you is important for you.
07:12:33 <xover> If you don't want that then its almost 50% overhead will always make it more expensive.
07:13:11 <Mike_L> nslu2 is only $80
07:14:42 <xover> The Drobo lets you buy drives — sans external cabinet — just-in-time (i.e. at a more advantageous point in the price curve) without major surgery (moving data between drives).
07:15:26 <xover> But as I said, it's expensive up front and its 50% RAID-ish overhead makes it hella expensive if you don't actually want the redundancy.
07:16:37 <tr__> nslu disco'd? damn
07:16:39 <Mike_L> so what kind of API should a backup service provide? I think it should allow the client to store objects consisting of a text ID, a blob of binary data, and various textual key-value pairs for metadata
07:17:50 <Mike_L> a simpler API would make a *single* text ID and a blob, and the app would embed all metadata into the blob. Then search and retrieval of blobs would require regex
07:18:07 <Mike_L> err the app would embed all metadata into the ID
07:18:22 <Mike_L> so the client could search for objects by providing a regex that would be matched against IDs
07:18:39 * tr__ wants FS with arbitrary metadata and fast querying
07:19:07 <xover> tr__: Spotlight?
07:19:17 <Mike_L> I don't know of any kind of index that makes abitrary regex run quickly
07:19:22 <tr__> xover: my macs run debian/gentoo ;)
07:19:36 <xover> Now see, there's you rproblem right there. :-)
07:19:53 <tr__> anwyays spotlight is orthogonal
07:20:02 <Mike_L> tr__: you're right that this backup service is kind of like a filesystem
07:20:03 <tr__> local spideringg / 3rd party index
07:20:29 <xover> But ext3 + xattr supports fairly arbitrary metadata, and there's some Spotlight-esque stuff out there for Linux. Beagle, maybe? I don't recall.
07:20:38 <Mike_L> the difference is that the backup service should be write-only read-many, and support only whole-object updates and retrieval
07:20:39 <tr__> yeah that xattr stuff.. never works for me, proably due to using reiser4 tho heh
07:21:00 <tr__> so yeah, i think just going HTTP makes sense, since you can use RDF for the metadata (or JSON, Yaml, whatever)
07:21:15 <Mike_L> tr__: HTTP for the API?
07:21:20 <tr__> Mike_L: what were you thinking?
07:21:27 <tr__> Mike_L: you wanted to backup blogs, etc right?
07:21:32 <xover> tr__: Are you familiar with the remedy usually prescribed when a patient says “Doctor, Doctor! It hurts when I do this”?
07:21:36 <Mike_L> any application, actually
07:22:19 <tr__> if a persistent/redundant /etc and ~/.* doesnt keep the app running right, it sucks ;)
07:22:27 <Mike_L> tr__: I wanted to define the requirements before considering the implementation
07:23:58 <xover> Hmm.
07:24:08 <Mike_L> I think that an application will store many different objects that all correspond to a single 'backup' and must all be retrieved if one is to restore the backup
07:24:41 <tr__> Mike_L: have you tried git?
07:24:42 <Mike_L> so it might be useful to represent this collection of objects
07:24:45 <sbp> hmm indeed
07:24:45 <phenny> sbp: 00:08Z <Morbus> tell sbp http://www.dragonrest.net/fortfiles/forthome.html
07:24:47 <Mike_L> tr__: yes
07:24:49 <sbp> dircproxy is still swhack-broken
07:24:51 <sbp> annoyings
07:24:53 <xover> You want to make a new backup service that's designed to enable the whole-app/app-managed backup that's possibly relvant for some specific apps?
07:25:09 <Mike_L> xover: yeah
07:25:13 *** tr__ is now known as a
07:25:20 <xover> And for this class of applications you think the backup service should be generic, but still specific for that class.
07:25:23 <Mike_L> I want to define the API and then protocol
07:25:35 <_bjoern> ----- _bjoern needs tool to make pretty ascii charts with multiple curves -----
07:25:58 <Mike_L> xover: I'm not sure
07:26:26 <xover> I would probably suggest identifying all the apps in the class, boil them down to a manageable number of representative examples, and then start making the specific requirements for each.
07:26:44 <sbp> phenny: tell Morbus I have the Dover edition that they mention. it is good
07:26:45 <phenny> sbp: I'll pass that on when Morbus is around.
07:27:03 <Mike_L> I think it would be great to allow the application or user to define a policy for the automatic delection of old backups, so we can use write-once and protect the data from malicious crackers
07:27:05 <xover> Once there you have a better foundation for figuring out the division of labour between the app's new backup facillity and the new backup service's API.
07:27:33 <_bjoern> ----- IS ANYBODY READING THIS -----
07:27:36 <Mike_L> _bjoern: yes
07:27:50 <_bjoern> THEN MAKE RECOMMENDATIONS PLZ
07:27:50 <xover> ----- NO -----
07:28:07 <xover> Elbow grease?
07:28:16 <Mike_L> _bjoern: does gnuplot have an ascii output driver?
07:29:08 <xover> .g "ascii output driver"
07:29:10 <phenny> xover: http://www.gnu.org/software/pspp/news.html
07:29:59 <_bjoern> "PSPP is a program for statistical analysis of sampled data. It interprets commands in the SPSS language and produces tabular output in ASCII, PostScript, or HTML format."
07:30:02 <_bjoern> nochartz
07:30:03 <Mike_L> xover: do you think I'm going to miss some important requirement?
07:30:26 <_bjoern> I was pondering making a png and throwing one of those png2asc toolz on it
07:30:38 <Mike_L> _bjoern: that might work
07:30:42 <_bjoern> likely I'll hate the result, so I didn't do it.
07:31:12 <xover> Well, yes; and no. I think you're thinking too abstractly/theoretically, which had the very real danger of reinventing RDF and the Semantic Web.
07:31:36 <_bjoern> gnuplot is claimed to support some ascii output
07:32:03 <_bjoern> Now I hate gnuplot a lot...
07:33:23 <Mike_L> xover: I don't see what this has to do with RDF or semantic web. RDF is for relaying information between applications. Backup is purely data store and retrieval by the same application (or different versions of it)
07:34:11 <Mike_L> _bjoern: gnuplot is easy to hate. It's badly in need of an interface upgrade
07:34:27 <xover> RDF and SemWeb have had an incredible amount of theoretical wankery applied; but so far has, for whatever reason, utterly failed to work in practice.
07:35:08 <_bjoern> lisppaste2: url?
07:35:08 <lisppaste2> To use the lisppaste bot, visit http://paste.lisp.org/new/swhack and enter your paste.
07:35:24 <lisppaste2> _bjoern pasted "gnuplot" at http://paste.lisp.org/display/61528
07:35:30 <xover> One reason for this is that they handwaved over the real world complication that most producers of data wouldn't publish it in a usable way even under physical duress.
07:35:50 <_bjoern> Now I need to make that suck less somehow...
07:36:22 <xover> You need to find out what similar snafus apply to your domain before you climb too far p the ivory tower (my metaphores are stretching, I know).
07:36:33 <a> i dont think you should poo-poop the idea
07:36:42 <a> of a backup service that provides a useful generalization for well written apps
07:36:52 <a> 96% of stuuff is always crap
07:37:03 <a> at least in software or music
07:37:31 * a likes the idea of redundant availablility more than backup
07:37:33 <xover> a: If I'm coming across as “poo-poop”ing the idea I must not be communicating clearly.
07:37:48 <a> power up my notebook, it realizes another one nearby has my approval to synch important stuff while theyre in proximity, etc
07:38:07 <xover> a: Now that sounds like a system I'd like!
07:39:03 <a> xover: app config is way more nasty than even the semantic web
07:39:09 <xover> Oh, also, it's some time very hard to distinguish between “redundancy” and “inefficiency”.
07:39:11 <a> xover: i mean, at least on web you jsut have atom, json, html, xml
07:39:15 <a> xover: every file in /etc/ is different :)
07:39:21 <a> its own serialization/parser usually
07:39:39 <xover> Yeah, the folks trying to reinvent the WIndows Registry on Linux have discovered that problem. :-)
07:40:36 <Mike_L> xover: I'll try to be more concrete
07:41:23 <xover> Mike_L: For reference, I'm throwing wrenches at your idea in the hopes that it'll help you succeede with it; not because I want you to give up or fail! (cf. a's comments above)
07:42:13 <thelsdj> wooo that doctor who was great, of course i had high expectations too
07:42:39 <Mike_L> I see two reasons to allow the application to group the stored objects/files together into a set: 1) it lets the application easily find all the necessary objects when restoring the backup, 2) it lets the data deletion policy identify groups of objects that should be deleted together
07:43:34 <xover> For instance, a Set might be and entire user blog, backed up from a multi-user hosted blog service?
07:44:15 <xover> Nested sets? config, templates, data are eacha set; and then collected in a meta-set of all three components?
07:44:17 <Mike_L> xover: a set would be a backup of the entire user blog at 2008-06-01 00:44 UTC
07:44:24 <xover> Or are you further down the stack here?
07:44:33 <Mike_L> I don't know if there would be much benefit for nested sets
07:45:44 <Mike_L> I guess the real benefit of the set would be to let apps identify the set without needing to look at the objects or interpret any app-specific metadata or blob data
07:45:53 <xover> How about I want a different backup service for my LiveJournal blog than another LiveJournal user?
07:46:05 <Mike_L> huh?
07:46:38 <xover> Do LiveJournal get to specify what backup service I use?
07:46:42 <Mike_L> I don't care
07:46:53 <Mike_L> that's not pertinent to the backup API
07:47:00 <Mike_L> backup service API
07:47:21 <xover> What if there are dependencies between the per-blog specific stuff and the setup of the hosting service?
07:47:26 <_bjoern> It should try harder to use, say, ` , ´ ... to smoothen the output but there seems to be no way to do that...
07:47:37 <_bjoern> plot with <style> has nothing suitable
07:48:05 <Mike_L> xover: that's an application design problem, not a backup service design problem
07:48:39 <Mike_L> xover: oh you mean the per-blog specific *configuration* and the hosting service *configuration* which will get backed up?
07:48:45 <xover> right
07:49:10 <Mike_L> I guess the application would need to sort those things out
07:49:43 <Mike_L> maybe the application would backup the per-blog config as one set, and the hosting-service config as another set?
07:50:24 <Mike_L> the app would just need to make sure that it can restore both sets when needed
07:51:05 <xover> So say we have Mike, Inc. hosting user blogs using GNU Blogging Server (GBS), and Xover, Inc. also hosting user blogs using GBS. What happens when you want to back up your blog from MikeBlog.om and restore it on blogs-by-xover.net?
07:51:11 <Mike_L> I guess that would bear on the problem of deleting old sets: one might need to store multiple kinds of sets and keep one of each... but this degenerates to nested sets
07:51:47 <Mike_L> I would prefer to have the application store a single set and manage its own nesting
07:52:02 *** cre8radix (n=cre8radi@p54BE5059.dip.t-dialin.net) has joined #swhack
07:52:12 <Mike_L> so when you go to restore the backup of 2008-06-01 00:44 UTC then the application lets you choose which parts to restore
07:52:35 <cre8radix> heya
07:52:43 <Mike_L> so I would reject nested sets as a necessary feature for a backup service API
07:53:03 <Mike_L> hi cre8radix
07:54:21 <Mike_L> how best to represent sets in the API? Should the application specify the set explicitly, or should the backup service choose the set based on date or session?
07:54:52 <xover> cf. VCS systems, and atomic comitts; perhaps?
07:55:45 <Mike_L> yeah, if we want atomic commits then we need a way for the application to signal the completion of the set. This would be implicit with session-based sets
07:56:38 <Mike_L> but session-based sets could prevent some network optimizations such as parallel puts of large objects, which might be necessary due to small TCP windows
07:57:20 <Mike_L> Can you think of any benefit of atomic commits for a backup service?
07:57:25 <xover> Applications designing around the TCP stack are generally a bad idea; it prevents later optimizations at the network level.
07:58:16 <xover> (sometimes it's *necessary*, but it's still a bad idea in general)
07:58:50 <Mike_L> I can't, so I would not include atomic commits at all, and also reject deriving the backup set from the session
07:59:31 <Mike_L> I guess the application will need to deal with the set ID upon restoring, since it needs to present the various sets to the user
07:59:34 <xover> SO how do you identify related blobs?
07:59:44 <Mike_L> so the application would be beter off choosing the set ID itself
08:00:14 <Mike_L> so this means each object needs at least two pieces of metadata: blob-ID, set-ID
08:00:19 <Mike_L> and I would add a timestamp, too
08:01:44 <Mike_L> does this seem reasonable?
08:03:33 <xover> Who generates the IDs, and how? How do you identify bit-wise identical blobs? How do you identify logically-indentical blobs?
08:03:48 <xover> Is the set-ID namespace global? Hierarchical?
08:04:21 <xover> .gc indentical
08:04:22 <phenny> indentical: 334,000
08:05:12 <Mike_L> the application generates the IDs
08:05:58 <Mike_L> it seems reasonable that the set-ID namespace is not hierarchical and is specific to the user-account used at the backup service
08:06:51 <xover> What if I have two applications using the same account at the backup service?
08:07:00 <Mike_L> now the question of bitwise and logical sameness is interesting
08:07:24 <xover> (cf. <http://tools.ietf.org/html/draft-ietf-usefor-message-id-01> btw)
08:07:44 <Mike_L> xover: then they had better name their sets differently. 'WordPress Backup of tamale.net/blog 2008-06-01 00:44 UTC' is a reasonable set-id
08:08:50 <xover> Consider localhost.localdomain. example.com.
08:09:09 <Mike_L> ?
08:09:28 <xover> See the I-D referenced above.
08:10:29 <xover> Your set-ID shares a lot of properties with USENET Message-IDs.
08:10:35 <Mike_L> I've often thought that a much more efficient backup system would have the application install an 'agent' in the backup server that would be able to detect logically identical objects
08:11:01 <xover> Now the backup service is no longer generic.
08:11:05 <Mike_L> xover: but set-IDs are limited to a single user
08:11:43 <Mike_L> xover: such a backup service could be generic if the agent were to run in a generic virtual machine and communicate with the data store and its application via an API
08:11:56 <Mike_L> but that's a long ways off
08:12:29 <xover> Yes, but the user isn't (or at least, shouldn't be) managing the IDs manually; how do you generate these automatically without creating collisions?
08:12:43 <Mike_L> why shouldn't the user manage her IDs?
08:13:56 <xover> Because the user will happily enter “Joe Blow's Blog” into both LiveJournal.com and WordPress.com.
08:14:18 <sbp> the user is always a her?
08:14:39 <xover> Why would you require the user to manage a “set-ID” in a “generic backup system”?
08:15:18 <Mike_L> when she getst he bill for her backup service she'll want to see whats in there.
08:15:39 * sbp finds generic female pronouns quite offensive, actually
08:15:48 <Mike_L> xover: hopefully the application wouldn't allow the user to make such a mistake
08:15:51 <sbp> when singular they is the correct choice
08:16:04 <xover> Mike_L: How?
08:16:19 <xover> sbp: Take your “They” and shove it up yer ass! :-)
08:16:20 <cre8radix> via dekaritae: http://elitaste.com/blog/2008/05/30/ladies-and-gentlementhe-mixtape-about-nothing/
08:16:23 <Mike_L> xover: the application chooses the set-ID. The user just consumes it
08:16:28 <cre8radix> nice mixtape
08:16:46 <sbp> xover: you'd say that to Shakespeare too I suppose?
08:16:54 <xover> Absolutely!
08:16:57 <sbp> heh, heh
08:18:14 *** kwijibo (n=kwijibo@162.Red-80-25-93.staticIP.rima-tde.net) has joined #swhack
08:18:14 <Monty> But what does kwijibo have to do with the price of fish?
08:18:46 <sbp> kwijibo!
08:18:53 <kwijibo> :D
08:18:55 <Mike_L> now the really interesting question is how to have the backup service automatically purge unneeded backups
08:19:05 <sbp> do I spy a Spanish ISP?
08:19:15 <sbp> also, time for news crawl
08:19:41 <sbp> not that I expect there to be any news
08:19:46 <sbp> given the Sundayness of the day
08:20:43 <sbp> [[[
08:20:43 <sbp> I recall, as the simplest instance, how he described a call paid at dusk on some neighbors at Rye, how he rang the bell and nothing happened, how he rang again and waited, how at the end there came steps in the passage and the foor was slowly opened, and there appeared in advance on the threshold, "Something black, something canine." To have said a black dog would not have done at all...
08:20:48 <sbp> ]]] - E. F. Benson on Henry James
08:21:16 <sbp> .gd foor
08:21:16 <phenny> foor: De kermis of foor is van oorsprong een jaarmarkt ter gelegenheid van de wijdingsdag van de patroonheilige van de plaats. ...
08:21:23 <sbp> eek make it stop
08:21:24 <Mike_L> for my own data, I would like to have the backup service keep my latest backup, one that's a week old, and one that's 3-months old
08:21:45 <cre8radix> http://www.ironsky.net/site/
08:21:49 <cre8radix> .title
08:21:52 <phenny> cre8radix: Iron Sky
08:21:56 <cre8radix> wow
08:22:13 <cre8radix> "In 1945 the Nazis fled to the moon. In 2018 they are coming back. "
08:22:58 <sbp> "...the first principle of civilization ought to have been, and ought still to be, that the condition of every person born into the world, after a state of civilization commences, ought not to be worse than if he had been born before that period."
08:23:05 <sbp> - Thomas Paine, via WL
08:23:08 <sbp> cre8radix: chuckle
08:23:39 <sbp> ahaha. the headline image on BBC News is this:
08:23:40 <sbp> http://newsimg.bbc.co.uk/media/images/44706000/jpg/_44706461_drinker_getty226i.jpg
08:23:50 <Mike_L> to guarantee the availability of 1-week and 3-month backups, the service would have to keep at least four backups... five including the latest backup
08:24:27 <kwijibo> sbp: your IP locating abilities are uncanny
08:25:11 <sbp> kwijibo: I'm a bit scared too!
08:25:12 <sbp> .title http://entertainment.slashdot.org/entertainment/08/06/01/0023244.shtml
08:25:14 <phenny> sbp: Slashdot | Pringles Can Designer Dies, Buried In a Pringles Can
08:26:25 <sbp> "CNN is reporting that videos from the Coachella music festival showing Prince covering Radiohead's 'Creep' have been removed by Prince's label, NPG records. Thom Yorke of Radiohead, when told of Prince's action, said 'Well, tell him to unblock it. It's our... song.' No comment from YouTube or Prince yet."
08:26:27 <sbp> - http://yro.slashdot.org/yro/08/05/31/1912204.shtml
08:27:37 <Mike_L> .g rima-tde
08:27:38 <phenny> Mike_L: http://openrbl.org/query?rima-tde.net
08:27:46 <sbp> .title http://news.bbc.co.uk/1/hi/world/7428476.stm
08:27:48 <phenny> sbp: BBC NEWS | World | Indian tribe found in Brazil
08:27:55 <sbp> - Indian? heh
08:28:11 <sbp> ...and that's the news
08:28:33 <a> http://www.taiwandaily.net/JPG/B01.jpg does it better
08:28:36 <a> the arrows make it
08:28:47 <sbp> yeah, we've all seen the story
08:28:53 <sbp> I was just amused at the title
08:29:12 <sbp> ooh, I hadn't seen that
08:29:14 <sbp> that's hilarious
08:29:47 <Mike_L> a: do you read chinese?
08:29:47 <sbp> the other thing I like about that front page is the smiling suns at the top
08:30:31 <a> nope Mike_L id like to learn but my todo list is pretty long already
08:31:06 * sbp also tries to calculate how fed up he is at seeing the American flag all the damn time
08:32:40 <sbp> it would be so bad if it didn't suck
08:32:45 <cre8radix> sbp: http://www.citybeat.de/news/Video-des-Tages-2008-05-29
08:32:46 <sbp> it should've been more like Cuba's, I think
08:32:54 <cre8radix> walk this way
08:33:12 <Mike_L> sbp: don't be petty
08:33:35 <a> its very offputting
08:33:42 <a> stars in the middle with some kind of radial pattern would be better
08:33:52 * a sort of cringes whenever seeing it
08:34:01 <Mike_L> a: you cringe at USA flag?
08:34:05 <a> yea
08:34:08 <Mike_L> why?
08:34:10 <a> theres mainly israeli flags around here tho
08:34:12 <a> cuz its ugly
08:34:18 <a> and , the whole blind-patriotism thing
08:34:35 <a> israeli flag is symmetrical, and theres no annoying red on it
08:34:42 <a> or ugly ass stripes
08:35:11 <Mike_L> I disagree about the USA flag
08:35:12 *** a is now known as carmen
08:35:32 <Mike_L> do you know what the parts of the flag represent?
08:35:39 <carmen> nope
08:35:48 <carmen> well the stars are states
08:35:51 <carmen> the stripes, 13 colonies?
08:35:58 <carmen> red: blood of the slaughtered indians?
08:36:35 <Mike_L> actually the blood of patriots who died fighting the British for USian independence
08:36:37 <carmen> if USA can have indians, why not Brazil?
08:37:21 <cre8radix> O_o
08:37:22 <Mike_L> every country has horrible shame in its history
08:37:44 <Mike_L> what matters is what we do today
08:37:45 <cre8radix> nations suck
08:37:54 <cre8radix> borders suck
08:38:21 <Mike_L> humans suck
08:38:29 <cre8radix> flags are for burninganimals too
08:38:46 <cre8radix> flags are for burnin / animals too
08:39:01 <Mike_L> cre8radix: flags are for animals?
08:39:26 <cre8radix> hrhr
08:39:43 * Mike_L tries to make sense of cre8radix's statement
08:39:46 <cre8radix> flags are for burning
08:39:59 <cre8radix> animals suck too
08:40:05 <sbp> cre8radix: oh man, that video's so fucked up
08:40:13 <cre8radix> cool, ey?
08:40:57 <sbp> Mike_L: well it could still have represented something interesting in a *different* design
08:41:10 <sbp> the idea is to design something which is meaningful *and* æsthetically pleasing
08:41:19 <sbp> it's got the meaningful bit down pat
08:41:41 <xover> I approve of anything symbolic of killing the British!
08:42:07 <sbp> even though the meaning is a retcon, isn't it?
08:42:13 <sbp> I thought it was taken from Washington's arms
08:42:36 * sbp grudgingly looks it up, knowing that he's going to have to see the damn thing again
08:42:37 <Mike_L> .ety retcon
08:42:37 <phenny> Can't find the etymology for "retcon". Try http://etymonline.com/?search=retcon
08:42:46 <xover> .wik Retcon
08:42:47 <Mike_L> .g retcon
08:42:49 <phenny> Mike_L: http://en.wikipedia.org/wiki/Retcon
08:42:51 <phenny> "Retroactive continuity (or informally retcon) is the deliberate changing of previously established facts in a work of serial fiction." - http://en.wikipedia.org/wiki/Retcon
08:43:12 <Mike_L> sbp: how is it a retcon?
08:43:17 <sbp> ahahaha
08:43:20 <sbp> it's even better than I thought
08:43:27 <sbp> it's based on the British East India Company flag!
08:43:31 <xover> heh heh
08:44:25 <sbp> Mike_L: well rather than use the real story of "we based this on Washington's arms/East India Co. flag", they made up shit about the blood of patriots and tacked on the idea about the number of states
08:44:35 <sbp> seems like a retcon to me
08:44:45 <sbp> retpatcon, perhaps
08:45:01 <cre8radix> http://youtube.com/watch?v=pSLYBz7rqYY&feature=related
08:45:14 <cre8radix> flag this
08:45:36 <Mike_L> "The origins of the design are unclear. It closely resembles the British East India Company (BEIC) flag of the same era, and an argument dating to Sir Charles Fawcett in 1937 holds that the BEIC flag indeed inspired the design.[14] However, the BEIC flag could have from 9 to 13 stripes, and was not allowed to be flown outside the Indian Ocean.[15] Both flags could have been easily constructed by adding white stripes to a British Red Ensign, a common fl
08:45:51 <sbp> yes, I read that too
08:46:58 <Mike_L> sbp: the new country desperately needed ideas to unify it. I think the people were smart to find unifying ideas to assign meaning to the new flag.
08:47:18 <sbp> you sure they didn't come later?
08:47:33 <sbp> [[[
08:47:33 <sbp> Given the scant archaeological and written evidence, it is unknown which design was the most popular at that time.
08:47:33 <sbp> The origin of the stars and stripes design cannot be fully documented.
08:47:34 <sbp> ]]]
08:47:35 <Mike_L> it doesn't matter to me
08:47:50 <sbp> but the point you're making depends on it!
08:47:53 <xover> Man. Rickrolling $cientology is a stroke of genious.
08:47:54 <Mike_L> your mockery comes across as petty
08:48:06 <sbp> mockery? I'm not mocking the flag
08:48:15 <sbp> I'm saying two distinct and fairly mundane things:
08:48:28 <sbp> 1) I see it way more often than any other flag in the world, probably even my own, which strikes me as odd
08:48:41 <sbp> 2) I don't like its design; I think it's one of the worst designs of country flags
08:48:52 <sbp> when you compound the two points, they add up to an annoyance
08:49:01 <sbp> I don't see where the mocking comes in. what am I mocking?
08:49:24 <cre8radix> xover: word!
08:49:37 <sbp> (I don't think my own country's flag is particularly good, note. it's a bit better than the American one, but probably in the bottom third or even quarter)
08:50:16 <sbp> (but even if I considered it the best, what then? it's just my æsthetic opinion)
08:50:25 <Mike_L> sbp: do you suppose people got tired of seeing the British flags during the heyday of the British empire?
08:50:35 <sbp> oh yes, I would guess very much so
08:50:48 <sbp> quite the symbol of evil back then
08:50:52 <sbp> like the American flag now, perhaps?
08:51:05 <sbp> that wouldn't change its æstheticality though
08:51:13 <sbp> it'd still be bottom third or bottom quarter I think
08:51:15 <Mike_L> I don't understand why you think the US flag is ugly. I like it.
08:51:44 <sbp> it looks like a toy flag... heraldically trite I guess one ought to say
08:52:02 <sbp> as I say, Cuba is quite good and quite similar
08:52:09 <Mike_L> blind patriotism irks me, too. A little informed patriotism is ok though
08:52:09 <sbp> note how it has fewer elements to achieve a stronger effect
08:52:43 <sbp> that's one of the greatest challenges in vexology: to come up with a design that is as simple and distinctive as possible
08:53:09 <sbp> adding states all the damn time didn't really help, so I admit it's partly due to historical accident
08:53:12 <Mike_L> the angle probably makes it more difficult to manufacture
08:53:21 <sbp> but that doesn't change the fact that spangling with stars looks kinda naff
08:53:26 <sbp> angle?
08:53:32 <Mike_L> the angles in the cuban flag
08:53:52 <sbp> oh, I see. er... dunno. how much does it cost to sew on all those stars? :-)
08:53:53 <Mike_L> the USA is a very large nation, made up of many states. It seems natural that each state should be represented.
08:54:15 <sbp> heh, again, that's counter to the received notions of what makes a good flag, as far as I'm aware
08:54:26 <Mike_L> I suppose the EU flag has a number of stars already
08:54:32 <sbp> you could have just had a whacking great heraldic eagle on it, for example
08:54:45 <sbp> that's a bold American symbol. it'd look better
08:55:23 * sbp doesn't like the EU flag, but again, bottom third but better than America
08:55:26 <sbp> at least it's simple
08:55:41 <sbp> though again does have that capacity to go nuts if we ended up with lots of splintered states
08:55:50 <sbp> a big ring of thirty stars would look silly
08:56:21 * sbp doesn't really expect any good design from the EU though, heh
08:56:31 <sbp> I'm more surprised about early America
08:56:37 <sbp> they were generally very sensible about such things
08:56:45 <Mike_L> I think you're trolling
08:56:50 <sbp> damn good nation builders, mythologisers, etc.
08:56:52 <sbp> I'm not!
08:57:24 <sbp> I mean, perhaps you're oversensitive because of generalised hostility to America because of its foreign policy?
08:57:35 <sbp> I know it gets lots of stray criticism that it doesn't deserve now
08:57:45 <sbp> but really mang, I reserve the right to say its flag is balls
08:58:47 <Mike_L> and do you like the UK flag?
08:58:54 <sbp> if a man can't say "the design of such and such a flag is not æsthetically pleasing" without raising a shitstorm, then I just dunno
08:59:02 <sbp> I already mentioned that I don't like it particularly, no
08:59:08 <sbp> 09:49 <sbp> (I don't think my own country's flag is particularly good, note. it's a bit better than the American one, but probably in the bottom third or even quarter)
08:59:08 <sbp> 09:49 <sbp> (but even if I considered it the best, what then? it's just my æsthetic opinion)
09:00:06 <sbp> since it's a bit better designed than the American flag though, and since *I see it less even though it's my own nation's flag*, it's less annoying in general to me
09:00:26 <Mike_L> the small diagonal red stripes on the UK flag seem misplaced... like the star pattern is twisted counter-clockwise
09:00:28 <sbp> also I made this as something of an aside comment from seeing it at the top of that Mandarin language newspaper
09:00:37 <sbp> yes, it's nuts
09:00:48 <sbp> I saw some interesting alternative designs somewhere
09:00:51 * sbp digs them out
09:00:57 <Mike_L> I'm not actually interested
09:01:16 <sbp> here we go: http://flagspot.net/flags/gb!alt.html
09:01:26 <sbp> well, also I hope you're not just fishing for reassurance
09:01:37 <sbp> because it could have turned out by chance that the UK flag was awesome
09:01:42 <sbp> it just so happened that that's not the case
09:01:45 <Mike_L> it's unfortunate that so many flags have religious symbols
09:01:47 <sbp> but if it were, as I say, then what?
09:02:05 <sbp> well, they're mostly quite old...
09:03:02 <sbp> secular states are kinda new as far as I'm aware, historically speaking
09:03:51 <Mike_L> before USA, what was the first secular state? Holland?
09:04:17 <sbp> good question. I don't know
09:04:21 <Mike_L> I suppose there have been many in ancient history...
09:04:26 <sbp> yeah, I was wondering that
09:04:37 <sbp> depends how you define it and such too
09:06:10 <sbp> what do you think the framers would've made of America now?
09:08:48 <Mike_L> I don't know. They were all very rich men. Most had plantations and employed slave labor. I think, once they got used to modern technology, they would be quite happy with modern USA
09:09:44 <sbp> well there are many things that would conflict with their ideology... but I've wondered if they just wouldn't care either way particularly
09:09:52 <xover> The USA is a secular state? When did this happen?
09:09:55 <sbp> I mean, they did what was right for *their* time, and now our time is different
09:10:00 <sbp> xover: it was originally
09:10:12 <sbp> certainly not now, though ostensibly so
09:10:30 <sbp> most of the framers were deists, I recall
09:10:31 <xover> Are they even trying to pretend to be?
09:10:46 <Mike_L> I hope they would be troubled by the imbalance of power in the federal government, with the executive branch claiming many powers
09:10:54 <sbp> well yes, I think they "pretend" to be out of reverence for the constitution and the days of the framers
09:11:11 <sbp> Mike_L: yeah, that's probably one of the big things...
09:11:23 <sbp> I mean, that's pissing on their system something chronic
09:11:41 <Mike_L> USA is still very much a secular state. Politicians pretend to be religious to get votes
09:11:41 <xover> I wonder when they'll have a Muslem president.
09:12:46 <sbp> I'd characterise it as deeply in the mire of quasi-secular, I think
09:12:48 *** chris2 (n=chris@p5B16BB18.dip0.t-ipconnect.de) has joined #swhack
09:13:07 <sbp> like Britain was a few decades ago
09:13:17 <sbp> when the church still had influence even though technically it didn't
09:13:22 <Mike_L> I think the framers would be concerned about the existence of the public primary and secondary school system
09:13:30 <sbp> (because of the class system)
09:13:36 <sbp> oh?
09:14:15 <Mike_L> yes, modern schools would be very strange to an early USian
09:14:25 <sbp> how so? I don't know anything about this
09:14:48 <Mike_L> and they would immediately question the need for huge bureaucracies to administer these bloated institutions
09:15:21 <xover> AH, and adherent of the Small Government Doctrine.
09:15:25 <xover> *an
09:15:47 <sbp> Jefferson could've taken on the enemy with his bare hands. no need for a Department of Homeland Security!
09:17:29 <Mike_L> xover: I don't know if my doctrine is 'small government'. I just think that it would never occur to a framer in the 1790s to create an institution to manage the one-room schools of all the towns
09:17:45 <Mike_L> since the parents seemed to do a decent job of that
09:18:37 <Mike_L> and I think a lot of them would totally oppose the Federal government trying to claim any authority over their local primary and secondary schools
09:19:00 <sbp> oh, another add-on point about flag design: it seems a shame when countries, which are the largest legal entities, use complex designs because they have the strongest right to claim the awesome really simple designs that most other entities can't really go for
09:19:37 <sbp> everybody knows the flags of Austria or France or Finland for example. so simple, but so recognisable
09:20:11 <cre8radix> well
09:20:16 <cre8radix> mine is black
09:20:20 <thelsdj> are there any flags that are just a single color?
09:20:20 <cre8radix> :D
09:20:41 <xover> The Parley flag.
09:20:47 <sbp> (though you do get problems of overlap then. like Italy and Ireland. or, worst of all, Netherlands and Luxembourg)
09:20:48 <Mike_L> sbp: flags consisting of bars of colors are hard to tell apart
09:20:55 <thelsdj> also, some flags have accessibility issues (colorblind)
09:21:25 <Mike_L> USA is geometrically distinctive
09:21:28 * sbp quickly scans the list of flags
09:21:41 <sbp> yes, Libya is just a field
09:21:41 <xover> Countries are the largest entities? This from the world's traditionally most accomplished Imperialist?
09:21:48 <cre8radix> phenny: de "Barrierefreie Fahnen"
09:21:59 <sbp> Libya's flag: “Vert.”
09:22:03 <cre8radix> phenny: de "Barrierefreie Fahnen"?
09:22:04 <phenny> cre8radix: "Barrierefreie flags" (de to en, translate.google.com)
09:22:21 <cre8radix> hrhr
09:22:22 <thelsdj> nice
09:22:30 <cre8radix> don't google me, phenny
09:22:31 <sbp> xover: individual countries these days have more power than any bloc
09:23:07 <sbp> Mike_L: geometrically distinctive with, say, Liberia?
09:23:15 <sbp> http://en.wikipedia.org/wiki/Image:Flag_of_Liberia.svg
09:23:34 <Mike_L> sbp: liberia copied usa
09:23:42 <sbp> I know :-)
09:25:34 <sbp> some of these young nations' flags are actually pretty good
09:25:43 <sbp> I like Cape Verde
09:26:09 <sbp> the Dominican Republic one is strangely compelling, too
09:28:52 <sbp> “The United States Army Institute of Heraldry has plans for flags with up to 56 stars, using a similar staggered star arrangement should additional states accede.”
09:28:59 <sbp> the circular 51 arrangement is freaky
09:37:11 *** kwijibo has quit ()
09:37:24 <sbp> hmm, interesting:
09:37:25 <sbp> http://en.wikipedia.org/wiki/Pledge_of_Allegiance#Addition_of_the_words_.22under_God.22
09:38:42 <xover> SS,CDL: Book #300 entered into my DL. :-)
09:38:43 <Mike_L> I don't like the idea of reciting an oath of allegiance
09:39:46 <Mike_L> it's ok for people entering public service: members of the military, leaders, representatives, and officers
09:40:24 <Mike_L> but it should be an oath of allegiance to the country, not a symbol of the country
09:41:19 <sbp> xover: yay :-)
09:41:45 <sbp> thelsdj: didn't you relay a story about some lady in Oregon(?) who refused to sign some documentation that pledged her allegience to God or something?
09:41:54 <sbp> any recollection of that?
09:42:52 <cre8radix> am i the only one here who doesn't give a shit about his country?
09:43:44 <sbp> oh, I think this might be one:
09:43:51 <sbp> er, this might be it even:
09:43:52 <sbp> .title http://www.outsidethebeltway.com/archives/2008/05/professor_refuses_to_sign_loyalty_oath/
09:43:53 <phenny> sbp: Professor Refuses to Sign Loyalty Oath » Outside The Beltway | OTB
09:43:59 <sbp> cre8radix: depends what you mean by giving a shit
09:44:11 <_bjoern> Today I learned that in XML 1.1 you can redeclare a prefix using xmlns:example='', but after you do that, using that prefix becomes illegal.
09:44:23 <sbp> it was on pacifism grounds
09:44:35 <sbp> _bjoern: heh
09:45:05 <_bjoern> You know that's the only new feature in Namespaces in XML 1.1, a 25 page recommendation in its second edition now.
09:45:26 <_bjoern> besides noting that xml 1.1 allows more chars in names than xml 1.0 of course.
09:45:35 <_bjoern> well and weird iri talk.
09:45:37 <Mike_L> bloat
09:46:40 <sbp> “Now, she would like to see the oath eliminated for all public employees except those who deal with sensitive information. She also would like an apology and a job next year.”
09:46:47 <sbp> doesn't look like that's happened, as far as I can find
09:46:56 <sbp> though I suppose it'd be less newsworthy if they did, so probably harder to find
09:47:19 *** cre8radix|off (n=cre8radi@p54BE5059.dip.t-dialin.net) has joined #swhack
09:47:58 * sbp searches more carefully
09:48:03 *** Mike_L is now known as Mike_L|off
09:48:07 <Mike_L|off> testing....
09:48:18 *** Mike_L|off is now known as Mike_L
09:48:52 <sbp> hmm, it seems to have happened more recently than I remembered
09:48:59 <sbp> so there's unlikely to be any update yet anyway
09:49:05 <Mike_L> 09:48:07 <Mike_L|off> testing....
09:50:34 *** carmen has quit ("leaving")
09:51:37 <sbp> [[[
09:51:38 <sbp> Gonaver is not the only instructor who lost a job over the oath. In February, Cal State East Bay fired a Quaker math instructor who signed the oath but inserted the word "nonviolently." She was rehired after Atty. Gen. Jerry Brown's office drafted a statement saying the oath would not commit her to bear arms.
09:51:43 <sbp> ]]] - http://www.latimes.com/news/local/la-me-oath9-2008may09,0,3786001.story
09:51:57 <sbp> Mike_L: cre8radix uses that suffix as |away, I think
09:52:07 <Mike_L> ok
09:52:13 <sbp> i.e. offline, I would suppose
09:52:27 <sbp> or off in the general sense as when one says “I'm off!”
09:55:35 <xover> …my rocker.
09:55:52 <xover> …my food.
09:56:28 <sbp> …ington Offlethwick of Offlestead Hall
09:57:28 * xover opts for “V for Vendetta” over the Kurosawa marathon, in honor of Anonymous.
09:57:41 *** Jibbler has quit ("Connection reset by Chuck Norris")
09:57:54 *** Jibbler (n=Jibbler@jibble.plus.com) has joined #swhack
09:57:57 *** libby (n=libby@92-237-82-160.cable.ubr17.azte.blueyonder.co.uk) has joined #swhack
09:58:07 <Mike_L> "Connection reset by Chuck Norris" <-- HAHA
10:03:27 *** cre8radix has quit (Read error: 113 (No route to host))
10:05:52 <deltab> .title http://news.bbc.co.uk/1/hi/england/london/7429638.stm — noun or verb?
10:05:54 <phenny> deltab: BBC NEWS | England | London | Tube drinks party sparks mayhem
10:06:53 <sbp> heh, d'oh
10:07:52 <_bjoern> hmm I think I found a bug in Apache's `ab`; seems you can't mix -k (persistent connections) and -i (use HEAD)
10:09:47 <sbp> deltab: Countryfile, btwz
10:09:59 <sbp> talking about bees dying out at the moment
10:10:29 <sbp> they think because of the weather and pests and diseases
10:10:43 *** mahound has quit (Connection timed out)
10:11:37 <deltab> ah, I heard about this in Doctor Who
10:12:08 <sbp> ...hehe
10:12:38 <deltab> you've seen that, right?
10:14:27 <deltab> subtitles: "... put a Tehran responsibiliter on the back ..."
10:14:44 <sbp> nope?
10:14:55 <sbp> I haven't watched Doctor Who since Christmas
10:15:14 <deltab> aw
10:15:52 <deltab> why not?
10:16:59 <libby> it was great last night
10:18:11 <sbp> each time it started up and I watched it, I didn't like it
10:18:20 <sbp> though I did really like the Christmas episode
10:20:43 *** cre8radix|off is now known as cre8radix
10:21:09 <cre8radix> _bjoern: http://www.youtube.com/watch?v=-HOUBpKXz64
10:21:12 <cre8radix> rofl
10:26:42 <deltab> sbp: why not?
10:34:15 <thelsdj> sbp: last nights was steven moffat, worth watching just for that
10:34:39 <_bjoern> good .title
10:35:02 <sbp> it's probably just that I don't like sci-fi much
10:35:14 <sbp> just a personal preference. it has to be really good before I watch it
10:35:29 <sbp> like, I'd watch a programme about knitting if it was *really* good
10:35:38 <sbp> but otherwise wouldn't normally
10:35:46 <thelsdj> yea, but like i said its steven moffat, that overrides all other considerations
10:36:01 <sbp> hehe
10:36:12 <sbp> what was its name?
10:36:22 <thelsdj> silence in the library (i think)
10:36:47 <_bjoern> "Sie steigen in den Hauptbahnhof ein!"
10:36:54 <sbp> I need something which gives me alerts about when particular episodes of things are on the telly
10:45:06 <deltab> iirc, Kaffe Fassett made some good tv about knitting a decade ago
10:45:32 <Mike_L> goodnight folks
10:46:45 <deltab> (or was it needlepoint?)
10:49:45 *** cre8radix has quit ()
10:50:34 <sbp> “it's the otter's very own fish and chip shop! ...without the fish” - Ben Fogle
10:50:52 <deltab> heh
10:54:48 <deltab> are you a fan of Agatha Christie at all? The episode a fortnight ago was about her ten-day disappearance
10:56:32 <deltab> and it had probably the funniest scene
10:56:58 <thelsdj> which scene? the parlor room?
10:57:22 <deltab> the cyanide antidote
10:57:28 <thelsdj> oh yes
11:00:13 <sbp> never watched it, I think
11:00:18 <sbp> would probably enjoy it
11:01:08 * sbp tries to fix Transmission
11:03:13 <sbp> 1.11 works
11:03:18 <sbp> 1.20 and 1.21 do not
11:09:16 *** mahound (n=mahound@unaffiliated/mahound) has joined #swhack
11:09:26 <thelsdj> Finding Neverland is good
11:15:50 <sbp> phenny: "Sie steigen in den Hauptbahnhof ein!"?
11:15:53 <phenny> sbp: "Get off at the main train station!" (de to en, translate.google.com)
11:21:07 <_bjoern> entering the main train station, like you'd enter a car.
11:21:55 <_bjoern> .gc blickwinkelkonzentrator
11:21:56 <phenny> blickwinkelkonzentrator: 13
11:33:49 *** kpreid has quit ()
11:35:04 *** kpreid (n=kpreid@cpe-69-202-162-8.twcny.res.rr.com) has joined #swhack
11:43:00 <_bjoern> http://www.youtube.com/watch?v=BEXPuoEN-uI
11:46:50 <sbp> .title
11:46:51 <phenny> sbp: YouTube - Stupid Horny Dog having Fun with a child Girl
11:48:06 <sbp> dogs are humptastic
12:11:33 *** eel (n=PircBot@cpe-69-202-162-8.twcny.res.rr.com) has joined #swhack
12:20:26 *** xover has quit ("Leaving")
12:35:38 *** swhask (n=swhask@cpe-69-202-162-8.twcny.res.rr.com) has joined #swhack
13:20:42 *** xover (n=xover@octet.neutri.no) has joined #swhack
13:21:42 <xover> Ah, that's much nicer.
13:22:52 *** besherman (n=jircii@red1c.ataco.se) has joined #swhack
13:22:53 <xover> The old 48-port monster 10/100 switch that was generating amount the same noise as a small passenger jet taking off is not replaced with a nice little 8-port fanless number, giving me 10/100/1000 to boot.
13:23:09 <xover> *now, even.
13:23:53 <xover> A wheelbarrow's worth of cables taken down, so I can even see the floor under my desk.
13:24:25 <xover> And an amount of dust great enough that you could probably judge its weight by heft has been excised.
13:25:41 <xover> It may just be my imagination, but with the VGA KVM switch gone, and DVI directly to the displays, even the el cheapo Dell things look much crisper.
13:29:21 <xover> Ah, but the 10.5.3 update doesn't seem to have had an appreciable effect on Delicious Library's scrolling; still feels pretty sluggish.
13:30:22 <deltab> I've seen noticeable fringing due to a KVM
13:30:48 <deltab> even compared to a direct analogue signal
13:32:23 <xover> This was a Dr. Bott MoniSwitch Dual VGA; which when I first got it didn't exhibit any degradation compared with direct VGA (even at 1600x1200, which was larger than average at the time).
13:32:44 <xover> Unless I'm imagining it, I'm inclined to attribute it to VGA vs. DVI.
13:33:09 <xover> Then again, I'm terrible at judging these things so it's likely I'm imagining it. :-)
13:47:33 <deltab> with an analogue connection you've got D to A to D conversion
13:47:42 <xover> right
14:03:25 *** [chr0n0s] (n=[chr0n0s@123.237.3.195) has joined #swhack
14:06:23 <sbp> http://www.numberspiral.com/
14:15:22 *** _bjoern has quit ("Quit")
14:16:14 <sbp> hello [chr0n0s]
14:16:24 <[chr0n0s]> bot?
14:16:31 <sbp> nope
14:16:44 <sbp> why would a bot reply over ten minutes after you joined?
14:16:55 <sbp> this is a publically logged channel
14:17:00 <sbp> please state your purpose
14:17:16 <[chr0n0s]> trolling around
14:17:18 *** libby has quit (Read error: 113 (No route to host))
14:17:21 <sbp> hehe
14:17:33 <sbp> this I support
14:17:39 <[chr0n0s]> heh
14:18:03 <[chr0n0s]> why are you asking the purpose?
14:18:15 <sbp> because this is a secret channel (+s)
14:18:24 <sbp> so people who come here usually have some specific thing in mind
14:19:09 <sbp> and sometimes if you don't ask 'em, they don't pipe up
14:19:19 <[chr0n0s]> oh... i found the website swhack.com and landed here
14:19:25 <sbp> aha
14:19:34 <sbp> welp, hellos
14:20:05 <sbp> only I am here
14:20:14 <sbp> everybody else has either gone fishin' or is in jail
14:20:22 <[chr0n0s]> well it's logged, not so secret, is it?
14:20:24 <[chr0n0s]> lol
14:20:31 <sbp> or, in a few cases, both
14:20:39 *** Arnia has quit ()
14:20:42 <sbp> yeah. more like semi-secret
14:20:45 *** libby (n=libby@92-237-82-160.cable.ubr17.azte.blueyonder.co.uk) has joined #swhack
14:20:59 <[chr0n0s]> !whoami
14:20:59 <swhask> Unknown command, try @list
14:21:04 <[chr0n0s]> @list
14:21:08 <sbp> but I mean random folk don't wander by from looking for channels with the word "hack" in them anymore
14:23:32 <[chr0n0s]> so swhack means software hack or something else?
14:23:48 <sbp> .g Swhack FAQ
14:23:50 <sbp> it's all in the FAQ my good mang
14:23:50 <phenny> sbp: http://swhack.com/faq/
14:26:40 <[chr0n0s]> What is Swhack? It's the cauldron of Morbus's fluids. The fuck?
14:26:41 <[chr0n0s]> !!
14:26:49 <sbp> mmm... fluids
14:28:49 *** _bjoern (n=bjoern@dslb-084-057-251-114.pools.arcor-ip.net) has joined #swhack
14:29:17 <[chr0n0s]> i am out, time to watch some cricket
14:29:23 <[chr0n0s]> the game
14:29:39 <_bjoern> yo hax teh cricket gamez
14:29:59 *** _bjoern changed the topic to: "Stallman, Bedos, Pava - the Unholy Trinity"
14:36:39 <nslater> Swhack: software hacking and leet sploits since 1999
14:37:05 *** jeffarch has quit (Read error: 110 (Connection timed out))
14:40:34 <kpreid> sbp: I thought the wanderers by were part of the fun?
14:46:12 <sbp> they seemed to be getting less humourless
14:46:16 <sbp> the problem with most of them is that they didn't respond much
14:46:25 <sbp> which just slaughters the comedy
14:47:57 *** Arnia (n=jgeldart@0-16-cb-bd-57-e.it.wlan.dur.ac.uk) has joined #swhack
14:48:49 *** libby has quit (Read error: 113 (No route to host))
15:06:23 *** idickinson (n=ijd@88-108-205-162.dynamic.dsl.as9105.com) has joined #swhack
15:10:57 <xover> *sigh* WHy oh why can't Safari have a decent plugin API…
15:22:00 <sbp> or at least a duckslot
15:22:23 <clsn> sbp: duckslot sounds... somehow obscene.
15:22:25 <nslater> do you think it would be worth my effort to push for @profile in html5, has anyone else already done this, where should I look etc etc
15:22:38 <sbp> clsn: it's a longstanding Swhack meme!
15:22:45 <sbp> nslater: I've done it, I think
15:22:50 <sbp> DanC has done it
15:22:50 <nslater> results?
15:22:57 <sbp> GUESS PLZ
15:23:06 <nslater> shit man, if DanC has done it I think I'll not bother
15:23:07 <Monty> potty mouth!
15:23:15 <sbp> I also proposed a profile link relationship
15:23:20 <nslater> I would like to read the results for mine own eyes
15:23:21 <sbp> which was quite funny
15:23:23 <clsn> cloaca: a hole lotta fun!
15:23:39 <clsn> .gd cloaca
15:23:40 <phenny> cloaca: The vent or common opening in birds through which the intestinal, urinary, and reproductive tracts empty.
15:23:50 <nslater> because as it currently stands, I refuse to use HTML5 while it doesnt have @profile
15:23:59 <nslater> not that anyone should care about that except me, but it is annoying
15:24:03 <xover> nslater: search the WG list for @profile. There are significant dollar-value-of-disks worth of email on the subject.
15:24:10 <sbp> heh, it's still on RelExtensions
15:24:22 <nslater> how can @profile be a relextension?
15:25:06 <sbp> <link rel="profile" href="http://example.com/yourprofile">
15:25:14 <nslater> aah, shweet
15:25:16 <sbp> then you can use other rel values per that profile
15:25:26 <nslater> which list, I have found this ammusing nugged:
15:25:28 <nslater> nugget
15:25:30 <nslater> [[[
15:25:34 <nslater> We have three lists:
15:25:34 <nslater> * A help list for Web designers * A discussion list for feedback on the specs * A place for implementors to compare notes * A mailing list for watching diff-by-diff commits to the spec
15:25:39 <nslater> ]]]
15:25:39 <nslater> THREE?
15:25:48 <nslater> http://www.whatwg.org/mailing-list
15:25:50 <sbp> :-)
15:26:03 <nslater> is this the w3c list I should be looking at or what?
15:26:10 <clsn> AMONG OUR LISTS...
15:26:13 <sbp> yes, www-html
15:26:15 <nslater> (I love how the topic has survived three days)
15:26:17 <sbp> er, not www-html
15:26:25 <sbp> w'ever the HTML WG list is
15:27:09 <nslater> yeah, used to be on that, argued for <initialism> or dropping <acronym> and adding @type to <abbr> and got laughed in to oblivion
15:27:12 <nslater> :(
15:27:31 <sbp> yes, I would have done too
15:27:46 <nslater> laughed at me?
15:27:50 <sbp> yeah
15:27:52 <nslater> why?
15:27:58 <sbp> because I mean the obvious thing to do is merge the two
15:28:04 <sbp> use their common prefix, say
15:28:08 <sbp> an <a> element
15:28:09 <nslater> example?
15:28:15 <nslater> thats what I proposed
15:28:16 <sbp> <a>W3C</a>
15:28:33 <nslater> <abbr type="initialism">HTML</abbr>
15:28:35 <nslater> for example
15:28:35 <clsn> (isn't there already an <a> element?
15:28:39 <clsn> )
15:28:42 * sbp waits f... thanks clsn
15:28:48 <clsn> Whew.
15:29:03 <nslater> is this it: http://lists.w3.org/Archives/Public/public-html/
15:29:09 <sbp> yes
15:29:20 <nslater> @profile returns two results :/
15:29:21 <sbp> by the way, I asked people about this
15:29:26 <sbp> and got two conflicting reports
15:29:29 <sbp> hang on
15:29:31 <nslater> about what?
15:30:11 <sbp> nslater: http://inamidst.com/whits/2008/02#profile
15:30:28 <nslater> yeah, seen that before
15:31:18 <nslater> what really annoyes me is not so much the lack of @profile but the part of the spec which says "you can use THESE rel values or rel values on the microformats wiki, but nothing else or it's invalid HTML"
15:31:28 <nslater> if they dropped the last part, I would be happy
15:31:31 <nslater> -ish
15:31:33 <sbp> http://www.w3.org/Search/Mail/Public/search?type-index=public-html&index-type=t&keywords=profile&search=Search
15:31:37 <sbp> 113 results
15:31:54 <nslater> difference between "@profile" and "profile" I guess
15:32:02 <sbp> including a thread going on quite recently
15:32:19 <sbp> [[[
15:32:22 <sbp> > 2) As a matter of fact, <http://microformats.org/wiki/profile-uris> states
15:32:22 <sbp> >
15:32:22 <sbp> > "In hcalendar-issues, it is ACCEPTED that each microformat should have a
15:32:22 <sbp> > profile URI, like the XFN profile (http://gmpg.org/xfn/11)."
15:32:22 <sbp> >
15:32:24 <sbp> > So it seems you are in disagreement with the microformats community as well.
15:32:26 <sbp> Actually the removal of profile="" was done in conjunction with the
15:32:28 <sbp> Microformats community. I do not know whether that page represents
15:32:30 <sbp> consensus in that community or not. I have asked for clarififcation.
15:32:32 <sbp> ]]] - Hixie, http://lists.w3.org/Archives/Public/public-html/2008May/0114
15:33:03 <nslater> [[[
15:33:05 <nslater> > > Summary: profile="" doesn't work in practice so we have dropped it. We
15:33:05 <nslater> > > haven't replaced it with anything since there isn't really a problem to
15:33:05 <nslater> > > solve -- conflicts don't occur in practice, as Microformat names are picked
15:33:05 <nslater> > > to be rather unique and recognisable. The market takes care of problems like
15:33:07 <nslater> > > this without the need for namespace syntax.
15:33:10 <nslater> ]]]
15:33:16 <nslater> the problem is REQUIRING microformats
15:33:49 <nslater> the whole reason microformats was/is "successful" is because they were allowed to do this kind of thing in the first place, by requiring microformats going forward they are preventing anything else similar from happening
15:34:03 <sbp> [[[
15:34:05 <sbp> GRDDL limiting itself to only apply to a subset of existing documents is a
15:34:05 <sbp> design decision that GRDDL has made, which is fine, but that shouldn't
15:34:05 <sbp> constrain the HTML5 design. HTML5 provides ways for GRDDL to hook into
15:34:05 <sbp> HTML5 (e.g. using <link rel="grrdl-profile">) if desired.
15:34:05 <sbp> In practice it isn't at all clear that GRDDL is in any way relevant in the
15:34:09 <sbp> real world (much like almost anything related to RDF), and so it isn't
15:34:13 <sbp> really an important concern for the development of the HTML language.
15:34:15 <sbp> ]]] - Hixie, http://lists.w3.org/Archives/Public/public-html/2008May/0102
15:34:29 <nslater> wow
15:36:25 <_bjoern> phenny, "Der sagt viel wenn der Tag lang ist."?
