00:00:30 <kate`> and was a little disconcerted to find that when he returned home, he'd turned into a giant insect
00:00:41 <calamari> kate`: and what alternative are you suggesting
00:01:08 <kate`> calamari, for this example, to store objects directly, instead of serialising them (files are serialisations of objects)
00:01:28 <kate`> calamari, that's not even a new idea, but it emphasises how stuck-in-a-rut we are in our ways
00:01:46 <kate`> the "it's always been this way" or "this works, why change it" train of thought disgusts me
00:01:55 <calamari> disks store bits in a serial fashion, even if they are random access
00:02:09 <kate`> yes, at a much lower level than we're talking about
00:02:31 <kate`> so does your ram: do you treat your structs as a string?
00:03:26 <kate`> imagine you have a jpeg: the first thing you do is parse it into a tree
00:03:34 <kate`> when you're done, you serialise that tree back to a file
00:03:47 <kate`> why not store the parsed tree in the first place?
00:04:05 <kate`> OS/400 (iirc) did this
00:04:16 <kate`> we seem to have forgotten about the idea
00:08:22 <kipple> hmm. interesting discussion :)
00:08:48 <kate`> if you think so, you may enjoy this: http://www.mca-ltd.com/martin/Ten15/introduction.html
00:09:02 <kipple> about the jpeg example: how is that a fault in the OS, and not the jpeg file format?
00:09:09 <SimonRC> (BTW, kate`, the answer is UNIX, with files-as-byte-streams. See the essay "The rise of Worse-is-Better")
00:09:31 <kate`> SimonRC, thanks :)
00:09:51 <kate`> kipple, it's the fault of the concept of a file existing at all: this is nothing specific for jpeg
00:10:06 <kate`> kipple, your XML document is not a bunch of ascii: it's a DOM tree
00:10:17 <kate`> kipple, ditto your haskell program
00:10:20 <SimonRC> see: smalltalk, with its "system image"
00:10:48 <kipple> but since the hardware is serial everything has to be stored serialized. the difference is just at what level the serialization occurs
00:11:03 <SimonRC> also, VMS, with its "indexed files" (database tables).
00:11:09 <kate`> kipple, yes, but this is a much lower level
00:11:35 <kate`> SimonRC, indeed :)
00:11:55 <kate`> SimonRC, i recognise this essay, btw
00:12:16 <SimonRC> some list terps saved and restored a memory image every weekend instead of doing run-time GC. :-D
00:13:08 * SimonRC really goes to bed this time.
00:14:29 <kipple> I'm not sure I completely understand what you want, but doesn't modern virtual machines like java and .NET do something like that?
00:15:29 <kate`> kipple, sure, a .NET OS would be equivalent. why has it taken since the mid 80s to appear?
00:15:43 <kate`> (it'd also be humonosly less beautiful, i bet)
00:16:07 <kate`> ten15 was around so much earler - and so were other beautiful examples
00:16:25 <kate`> the essay SimonRC mentioned explains why: "worse is better" has better survival characteristics
00:16:30 <kate`> i.e. they win because of marketing
00:19:52 <kate`> http://www.jwz.org/doc/worse-is-better.html
00:20:04 <kate`> "Once the virus has spread, there will be pressure to improve it, possibly by increasing its functionality closer to 90%, but users have already been conditioned to accept worse than the right thing."
00:20:29 <kate`> compare that to the ideas earlier for small incremental improvements (adding features, etc)
00:20:38 <kate`> and see how heavily ingrained that concept has become
01:01:07 -!- kipple has quit (Read error: 110 (Connection timed out)).
01:18:07 -!- GregorR has joined.
01:42:26 -!- calamari has quit ("Leaving").
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:03:19 -!- pierce_ has joined.
08:04:07 -!- pierce_ has left (?).
11:34:34 -!- kipple has joined.
16:03:03 <ihope_> No conversation in over TWELVE HOURS.
16:44:06 <kipple> that's rather common in this channel...
17:09:51 -!- nooga has joined.
17:22:51 <nooga> why egobot is not here?
17:54:15 <Baughn> His creator had an acute attack of humility
17:57:17 <Baughn> Saturn was in presentile conjunction with mars, causing fear of morbidity in all egobot creators.
18:00:05 <Baughn> kate`: ..what, do you want your palm read?
18:01:35 <kate`> there're people doing that in town - if the process didn't involve giving them money, i'd try it, just to see what they have to say
18:02:04 <kipple> a bit hard to read the palm over IRC...
18:02:18 <kate`> i could draw it in ascii art
18:02:42 <Baughn> Hah. A *true* palm reader can see your palm even across half a planet.
18:04:51 <kipple> hmm. an online palm reading service... that could be a good business plan :D
18:05:56 <kipple> "please place your palm on your mouse and sit still for 10 seconds while concentrating on positive energies"
18:06:28 <nooga> AAAAAA NOOOO! attack of the VIDEO-TAPE-RECORDER-MAN!!! ...again?!
18:08:29 <nooga> kipple: that's damn good idea, i think i'll implement it because i've got loads of free time
18:08:47 <Baughn> Mouse? "Please place your palm on the screen for ten seconds while we scan it. (See this visible moving line)"
18:09:20 <nooga> but if someone has an optical mouse
18:09:56 <nooga> with this terrifying red laser eye
18:11:09 * kate` goes to sort through bank stuff
18:12:04 <kipple> "for extra accurate palm readings please purchase our patented USB palm scanner for only $299.99"
18:13:03 <nooga> i bet that there is a group of pph who would be ready to buy it
18:23:59 <nooga> online church fete store? ;d
18:24:14 <nooga> or selling church fetes on ebay
18:25:14 <nooga> AFAIK it's a deletion of all your sins
18:25:40 <nooga> of them... depends on price ;d
18:26:37 <nooga> quite popular in middle ages
18:27:46 <SimonRC> put my list of grades through: plot . zip [1..] . map (log . (100-)) . sort and you get an almost straight line :-S
18:30:05 <SimonRC> except for the grade of 100%, which goes wrong
18:30:26 <SimonRC> heh: "But after a while I learned the trick of speaking fast. You don't have to think any faster; just use twice as many words to say everything."
19:26:49 -!- nooga has quit.
19:44:57 -!- calamari has joined.
20:03:00 -!- kipple has quit (Read error: 110 (Connection timed out)).
20:43:29 -!- calamari has quit (Read error: 110 (Connection timed out)).
20:43:34 -!- calamari has joined.
21:17:48 -!- GregorR_ has joined.
21:18:42 -!- GregorR has quit (Nick collision from services.).
21:19:22 -!- GregorR_ has changed nick to GregorR.
21:28:58 <calamari> I guess you're sleep chatting? :)
21:29:50 <ihope_> One could say that :-)
21:56:18 -!- sekhmet has quit ("Lost terminal").
21:56:34 -!- sekhmet has joined.
22:45:37 <SimonRC> I heard about a guy who would cook meals in his sleep
22:45:45 <SimonRC> He was a professional chef.
23:02:13 -!- ihope__ has joined.
23:02:22 -!- ihope__ has changed nick to ihope.
23:04:17 -!- graue has joined.
23:11:10 -!- ihope_ has quit (Connection timed out).
23:22:58 <calamari> got another interp down to 42 bytes.. was hoping for smaller
23:23:24 <lament> what does it interpret?
23:24:18 <graue> is that 42 bytes for the binary?
23:25:45 <calamari> graue: btw, curious if you implemented those spam changes that were posted
23:27:47 <calamari> I think the best solution, if code is changing, is to implement a password box.. then the browser can save the password
23:28:09 <calamari> I doubt very much if the spammers will bother to update their code for one wiki
23:28:36 <calamari> I did this on the qemu wiki and haven't had any spam since
23:29:31 <calamari> of course the password is printed right next to the box.. basically the simplest captcha possible
23:36:48 <ihope> So it would be a machine-readable CAPTCHA, pretty much?
23:37:28 <calamari> yes, but as long as no machine implements it, who cares
23:38:00 <calamari> another idea along the same lines is to change the html so the form is different
23:38:19 <calamari> wikipedia hides values in the form
23:38:39 <calamari> but this is slightly better I think, because it doesn't rely on browser tricks
23:39:46 <calamari> sorry, didn't explain that ery well, did I
23:41:35 <calamari> wikipedia sends a value with the editing page, whihc the browser automaically submits back to them, so they know it was a browser
23:51:09 <pgimeno> an even simpler captcha method is what I implemented in a phpBB forum a few weeks ago
23:51:25 <pgimeno> "Always check this checkbox: [ ]"
23:51:52 <pgimeno> http://lordkiri.lo.funpic.org/ (when trying to post)
23:52:22 <pgimeno> spam is no longer a problem there
23:53:55 <pgimeno> (besides, there's another change that is not relevant here: when registering a new user, no URL is allowed)
23:55:57 <pgimeno> I mean, the "Web page" field *must* be blank
23:56:08 <pgimeno> there were massive registrations of spammers
23:56:20 <pgimeno> to increase Google rating, I suppose
23:57:40 <pgimeno> of course I also changed the "Web page" title for the field to "Leave this blank"
23:57:43 <calamari> btw, did you have the confirm of email addresses in place?
23:57:54 <pgimeno> yes but spammers didn't care
23:58:16 <pgimeno> I suppose they just wanted the address to appear in the user list
23:58:27 <pgimeno> (the users appear in the list even if not active)
23:59:01 <pgimeno> either that, or they don't care whether the email activation feature is in effect