| 00:53.42 | *** join/#openmoko timtimred (n=On3@82-45-84-42.cable.ubr02.chel.blueyonder.co.uk) |
| 01:30.42 | Sum1 | Phew, email sent. Bed time now. |
| 01:31.01 | Sum1 | Night all. |
| 08:00.46 | *** join/#openmoko dullard (n=jim@adsl-static-1-30.uklinux.net) |
| 08:15.43 | *** join/#openmoko koen (n=koen@dominion.kabel.utwente.nl) |
| 09:01.08 | *** join/#openmoko mikesh (n=mikesh@x133.net.upc.cz) |
| 09:04.45 | *** join/#openmoko koen (n=koen@dominion.kabel.utwente.nl) |
| 09:09.49 | *** join/#openmoko Sum1 (n=Sum1@80.175.142.81) |
| 09:12.26 | *** join/#openmoko mikesh (n=mikesh@x133.net.upc.cz) |
| 10:42.10 | *** join/#openmoko den-ros (n=den@ppp91-76-78-255.pppoe.mtu-net.ru) |
| 10:47.19 | *** join/#openmoko XVampireX (n=serge@85.64.236.111.dynamic.barak-online.net) |
| 10:56.39 | *** join/#openmoko [nodus] (n=[nodus]@c-67-188-43-151.hsd1.ca.comcast.net) |
| 11:40.06 | *** join/#openmoko zed_w0rm (n=Rik_07@host188-212.pool8249.interbusiness.it) |
| 11:41.09 | zed_w0rm | zomebaody can tell me the dimension of openmoko? |
| 11:41.49 | zed_w0rm | and where to buy? |
| 11:41.53 | zed_w0rm | please |
| 11:49.02 | buz | cant buy it yet |
| 11:49.11 | buz | siez is somewhere on the site, go to press and scroll down |
| 11:50.03 | zed_w0rm | oki |
| 11:50.04 | zed_w0rm | tnx |
| 11:50.23 | zed_w0rm | and when can i buy? |
| 11:50.33 | hrw | zed_w0rm: nowhere |
| 11:50.45 | hrw | zed_w0rm: it will be released in January |
| 11:50.53 | zed_w0rm | oki |
| 11:50.56 | hrw | so then it will be available to buy |
| 11:51.06 | zed_w0rm | 350 $? |
| 11:51.08 | hrw | zed_w0rm: read openmoko-community mailing list archives |
| 11:51.30 | hrw | thats current price |
| 11:51.46 | hrw | how much it will on release we do not know |
| 11:51.49 | *** join/#openmoko marcb4242 (n=m@car06-3-82-240-156-91.fbx.proxad.net) |
| 11:52.14 | hrw | there will be developer discounts probably |
| 11:52.23 | hrw | I count on it ;D |
| 11:52.31 | zed_w0rm | ;D |
| 11:52.39 | zed_w0rm | I NEED LINUX |
| 11:52.49 | zed_w0rm | IN MY POCKET |
| 11:53.09 | hrw | zed_w0rm: do not shout |
| 11:53.19 | zed_w0rm | so i can compile also in the wc |
| 11:53.23 | hrw | and go buy pda supported with linux |
| 11:53.26 | hrw | hahaha |
| 11:53.47 | hrw | another 'linux == compilation' guy you are? |
| 11:54.13 | zed_w0rm | a little |
| 11:54.34 | zed_w0rm | what is the best pda for linux? |
| 11:54.44 | hrw | depends what you want etc. |
| 11:54.57 | hrw | ipaq h2210 is nicely supported one but lack wifi |
| 11:55.02 | zed_w0rm | gps and open source |
| 11:55.13 | hrw | internal gps? no way |
| 11:55.21 | hrw | external BT/serial gps - many choices |
| 11:55.39 | zed_w0rm | :/ |
| 11:56.29 | zed_w0rm | and other linux phones? |
| 12:09.31 | [nodus] | i have a zaurus |
| 12:09.35 | [nodus] | its nice |
| 12:09.55 | [nodus] | i would love if sone one made a gprs compact flash |
| 12:10.20 | [nodus] | I would buy one if they cost less than 150 |
| 12:11.39 | [nodus] | i googled a year ago and there was nothing |
| 12:11.44 | [nodus] | but i guess thats chainged |
| 12:11.50 | [nodus] | http://www.mobiletechreview.com/tips/Enfora_GSM_CF_card.htm |
| 12:15.45 | *** join/#openmoko Sum1 (n=Sum1@80.175.142.81) |
| 12:34.11 | *** join/#openmoko yure (n=yure@dhcp-212-235-229-194.sd.uni-lj.si) |
| 12:35.15 | *** part/#openmoko zed_w0rm (n=Rik_07@host188-212.pool8249.interbusiness.it) |
| 12:37.41 | buz | are there still pda form factor zaurus being made? as opposed to those tiny laptop ones |
| 12:38.15 | [nodus] | the 6000 is pretty common |
| 12:38.34 | [nodus] | the 5500 is what I have |
| 12:38.49 | [nodus] | its pretty cheap to buy used these days. |
| 12:47.11 | [nodus] | the kb works suprisingly well |
| 12:47.35 | *** join/#openmoko incinerator (n=sabine@82-41-24-164.cable.ubr04.edin.blueyonder.co.uk) |
| 13:28.16 | *** join/#openmoko buz (n=buz@84-73-95-139.dclient.hispeed.ch) |
| 14:28.16 | *** join/#openmoko koen (n=koen@dominion.kabel.utwente.nl) |
| 14:48.31 | *** join/#openmoko [g2] (n=g2@nslu2-linux/g2) |
| 15:49.04 | *** join/#openmoko _kaiser_ (n=kaiser@74.130.78.171) |
| 16:31.32 | *** join/#openmoko XVampireX (n=serge@85.64.236.111.dynamic.barak-online.net) |
| 16:51.21 | *** join/#openmoko XVampireX (n=serge@85.64.236.111.dynamic.barak-online.net) |
| 16:57.05 | *** join/#openmoko suihkulokki (i=rvoipio@nat/movial/x-b85a1a1fc193b2ae) |
| 17:18.25 | *** join/#openmoko fragggy (n=fragggy@dslb-084-062-182-037.pools.arcor-ip.net) |
| 17:27.40 | *** join/#openmoko lama (n=lama@estel.wpia.uw.edu.pl) |
| 17:47.48 | *** join/#openmoko markit (n=konversa@host119-245-static.72-81-b.business.telecomitalia.it) |
| 18:03.15 | *** part/#openmoko suihkulokki (i=rvoipio@nat/movial/x-b85a1a1fc193b2ae) |
| 18:11.59 | *** join/#openmoko cephdon_ (n=cephdon@c-24-21-183-69.hsd1.or.comcast.net) |
| 18:19.01 | *** join/#openmoko XVampireX_ (n=serge@89.1.243.238.dynamic.barak-online.net) |
| 18:54.33 | *** join/#openmoko koen (n=koen@dominion.kabel.utwente.nl) |
| 19:10.18 | *** join/#openmoko khatahn (n=khatahn@MDCLXXV.voas.saunalahti.fi) |
| 19:26.33 | *** join/#openmoko timtimred (n=On3@82-45-84-42.cable.ubr02.chel.blueyonder.co.uk) |
| 19:36.01 | *** join/#openmoko GakkGakk (n=gakkgakk@209.84-48-194.nextgentel.com) |
| 19:50.58 | *** part/#openmoko GakkGakk (n=gakkgakk@209.84-48-194.nextgentel.com) |
| 20:06.00 | *** join/#openmoko bedboi (n=bedboi@87.18.74.110) |
| 20:06.02 | bedboi | hi there |
| 20:09.33 | koen | hey signore DeCicco |
| 20:09.34 | bedboi | is there any way to get a neo? |
| 20:09.45 | bedboi | hey koen! |
| 20:09.48 | bedboi | what's up? |
| 20:10.18 | koen | not much right now |
| 20:11.54 | bedboi | koen: do you know if there is some possibility to get neo1973? |
| 20:12.09 | bedboi | we, as university, would like to do some research on that |
| 20:12.12 | koen | yes, wait till 2007 and order one :) |
| 20:12.36 | koen | otherwise see http://www.openmoko.com/ |
| 20:53.25 | timtimred | oo, mobile & embedded java is open sourced. |
| 20:54.03 | *** join/#openmoko mjr (i=mjrauhal@128.214.205.15) |
| 20:54.38 | koen | still slower as native code |
| 20:55.57 | timtimred | true but users like their j2me games and stuff. |
| 20:59.23 | Stephmw | yup, yup |
| 21:00.20 | Stephmw | for all its faults, j2me does give a tad more portability than native code |
| 21:00.30 | Stephmw | and security |
| 21:00.42 | timtimred | and its a really popular technology for phones. |
| 21:02.49 | hrw | and I have one j2me application which use each time when I go somewhere in city |
| 21:02.49 | hrw | bus/tram timeline |
| 21:07.20 | *** join/#openmoko yure (n=yure@dhcp-212-235-229-194.sd.uni-lj.si) |
| 21:08.47 | xorAxAx | hrw: metro? |
| 21:10.25 | hrw | xorAxAx: no - this application support only Poznan, Poland |
| 21:10.35 | hrw | metro is totally other kind of app |
| 21:10.35 | *** join/#openmoko markit (n=konversa@host119-245-static.72-81-b.business.telecomitalia.it) |
| 21:10.36 | xorAxAx | ah |
| 21:27.54 | *** join/#openmoko lkcl (n=lkcl@host81-157-198-228.range81-157.btcentralplus.com) |
| 21:41.58 | koen | heh |
| 21:42.10 | koen | wanna bet which excuse ARM ltd uses now? |
| 21:42.22 | koen | robtaylor: and I think you mean Thumb2EE |
| 21:43.19 | robtaylor | koen: well whoever did it would have to RE Jazelle ratehr than using official docs :) |
| 21:43.31 | *** join/#openmoko rwhitby (n=rwhitby@nslu2-linux/rwhitby) |
| 21:45.39 | robtaylor | koen: and i think i do mean PhoneME |
| 21:45.57 | robtaylor | but then again i don't know a whole lot about sun's stuff |
| 21:46.33 | koen | robtaylor: ARM renamed jazelle to thumb2EE |
| 21:47.03 | robtaylor | ohh, lol |
| 21:47.07 | robtaylor | missed that one :) |
| 21:47.21 | robtaylor | an ever crapper name, brilliant ;) |
| 21:51.10 | Stephmw | what? link please? |
| 22:06.28 | robtaylor | koen: ah, actually thumb2-ee is Jazelle RCT |
| 22:06.41 | robtaylor | i guess DBX is still called jazelle |
| 22:07.05 | koen | still marketing nonsense |
| 22:07.21 | koen | just like iwmmxt |
| 22:09.24 | Stephmw | makes you wonder... Thumb2EE to generic bytecode accelerator... including .net? |
| 22:11.10 | robtaylor | Stephmw: yeah, it'd probably be useful for .net |
| 22:11.24 | robtaylor | koen: 8nod* |
| 22:12.31 | koen | you heard about o-hand efforts to port kdrive fbmmx to iwmmx? |
| 22:13.07 | robtaylor | ooh, no, not yet |
| 22:13.24 | koen | it's not faster as without iwmmx |
| 22:13.40 | koen | maybe for xrender, but that makes it crash |
| 22:14.30 | robtaylor | lol |
| 22:15.16 | koen | X couldn't deliver 8 bit aligned pixmaps |
| 22:16.08 | xorAxAx | WTH is Thumb2EE? |
| 22:16.58 | xorAxAx | jazelle sounds like an ARM jit |
| 22:18.56 | robtaylor | xorAxAx: Jazelle is ARMs instruction set extention that executes java natively |
| 22:19.23 | xorAxAx | ah |
| 22:19.26 | robtaylor | xorAxAx: thumb2-EE is a small extention that.. |
| 22:19.49 | koen | 'is said to execute parts of bytecode natively' |
| 22:19.50 | xorAxAx | they decode the bytecode in hardware? |
| 22:19.55 | xorAxAx | i dont think so |
| 22:20.19 | koen | xorAxAx: yes, you can throw bytecode at it, and it will execute it |
| 22:20.33 | robtaylor | "New features provided by Thumb-2EE include automatic null pointer checks on every load and store instruction, an instruction to perform an array bounds check, and the ability to branch to handlers, which are small sections of frequently called code, commonly used to implement a feature of a high level language, such as allocating memory for a new object. |
| 22:20.34 | koen | a subset of bytecode last I heard |
| 22:20.38 | robtaylor | [edit] " |
| 22:21.09 | robtaylor | xorAxAx: yep, theres some instruction sthat need to be done in arm code, but a lot that justexecute directly |
| 22:21.14 | xorAxAx | koen: it? |
| 22:21.15 | robtaylor | there's a seperate processor mode |
| 22:22.14 | robtaylor | so basically, you take a jit, use thumb2-ee where it makes sense, and reduce the amount that gets compiled, depending on the instruction |
| 22:22.21 | robtaylor | and bob's your proverbial |
| 22:22.45 | koen | bring on the meta-puns |
| 22:23.06 | xorAxAx | i am reading the whitepaper curretly :) |
| 22:23.11 | robtaylor | heh |
| 22:23.38 | koen | I threw that away last week while sorting out the stack of papers next to my bed |
| 22:23.55 | robtaylor | koen: you still use *paper*?!! |
| 22:24.16 | koen | jazelle can <buzzword> the <buzzword> and <buzzword> <buzword> ... zzzzzzzzzzzzzz |
| 22:24.49 | xorAxAx | 12k gates, wow |
| 22:25.20 | koen | ARM used to say "we can't give you docs because java is under a sucky license" |
| 22:25.29 | robtaylor | I actually reckon it'd be pretty simple to modify some smallish jit to use jazelle |
| 22:25.32 | robtaylor | koen: lolol |
| 22:25.36 | koen | any guesses which lame excuse they pull out now? |
| 22:25.59 | robtaylor | koen: they're already used to giveing people restrictive NDA's for looking at their instruction sets |
| 22:26.15 | robtaylor | koen: see the quemu arm9/11 debacle.. |
| 22:26.17 | xorAxAx | koen: even if you bought their ARM design? |
| 22:26.21 | koen | which is why qemu can emulate arm11, right? |
| 22:26.23 | robtaylor | xorAxAx: yes |
| 22:26.36 | robtaylor | koen: it can now?! |
| 22:26.45 | koen | ehm |
| 22:26.46 | koen | can't |
| 22:27.01 | xorAxAx | and its not possible to read the code? are they just supplying binary intermediate representation? |
| 22:27.02 | robtaylor | indeed |
| 22:28.31 | robtaylor | xorAxAx: it'd be simple to RE jazelle itself, all you need is to figure out the instruction to switch modes |
| 22:28.45 | robtaylor | xorAxAx: thumb2-ee would be much harder |
| 22:28.57 | koen | odjdump -d <some app with jazelle> ? |
| 22:29.19 | xorAxAx | hmm, they have a funny benchmark in the whitepaper |
| 22:29.36 | robtaylor | koen: exactly |
| 22:29.44 | xorAxAx | it claims that the arm vm is nearly as fast as another "third party commerical VM" with appropriate support |
| 22:29.54 | xorAxAx | no, i mean, its just a bit faster |
| 22:30.04 | xorAxAx | so there is no real reason to buy the arm vm :) |
| 22:31.17 | *** join/#openmoko necrodome (n=chatzill@85.97.141.140) |
| 22:33.39 | necrodome | hi |
| 22:39.07 | robtaylor | http://sourceware.org/ml/binutils/2002-01/msg00373.html |
| 22:40.16 | xorAxAx | hehe |
| 22:40.16 | robtaylor | there you go, basic jazelle support sorted ;) |
| 22:40.39 | xorAxAx | is there java bytecode disasm support? :) |
| 22:40.41 | robtaylor | gas has support for BXJ as well |
| 22:40.55 | robtaylor | xorAxAx: eh? in what sense? |
| 22:41.15 | xorAxAx | so that you can disassemble the target where BXJ jumps to |
| 22:41.34 | robtaylor | oh, lol |
| 22:41.41 | robtaylor | it probably does tbh |
| 22:45.12 | xorAxAx | hmm, i just read that bender is powered by a 6502 |
| 22:45.48 | robtaylor | of course! |
| 23:07.17 | xorAxAx | "To all licensees, ARM provides an integratable hardware description of the ARM core, as well as complete software development toolset (compiler, debugger, SDK), and the right to sell manufactured silicon containing the ARM CPU." |
| 23:07.31 | xorAxAx | "More ambitious customers, including integrated device manufacturers (IDM) and foundry operators, choose to acquire the processor IP in synthesizable RTL (Verilog) form. " |
| 23:07.51 | xorAxAx | so you can actually check how they implemented their jazelle stuff |
| 23:08.07 | xorAxAx | but you wont be able to tell anybody without a law suite in the back |
| 23:08.13 | Stephmw | right |
| 23:08.20 | Stephmw | what a load of tosh |
| 23:10.24 | xorAxAx | "A single customer product containing a basic ARM core can incur a one time license fee in excess of (USD) $200,000. Where significant quantity and architectural modification are involved, the license fee can exceed $10M." |
| 23:10.40 | xorAxAx | wow, i should design a new RISC processor :) |
| 23:12.44 | xorAxAx | and then after i got a few customers, i would make it open source :) |
| 23:18.06 | xorAxAx | hmm, how many gates does a T1 have? |
| 23:18.37 | MDK | do you happen to know something about jazzele perf.? |
| 23:18.51 | MDK | from what I heard, it's much hussle about nothing |
| 23:18.55 | xorAxAx | MDK: there is a whitepaper ... |
| 23:19.22 | xorAxAx | http://www.arm.com/pdfs/JazelleWhitePaper.pdf |
| 23:19.25 | Stephmw | MDK: yeah, but I can't tell ya :) |
| 23:19.45 | xorAxAx | hmm, a T1 seems to have 16.5 million logic gates |
| 23:25.34 | *** join/#openmoko tapo (n=tklopp@dslb-084-060-054-020.pools.arcor-ip.net) |
| 23:25.44 | xorAxAx | MDK: factor 3 is not that much |
| 23:26.22 | MDK | xorAxAx: can hardly believe it's 3 |
| 23:26.30 | xorAxAx | me neither |
| 23:26.56 | xorAxAx | i guess memory savings are a bigger advantage, no need to keep chunks of native code in memory |
| 23:27.06 | MDK | but I heard that next version of arm is going to have cyclope |
| 23:27.10 | MDK | a hardware-accelerated C |
| 23:27.16 | MDK | has great improvements |
| 23:27.20 | xorAxAx | how should that work? |
| 23:27.25 | MDK | especially, when you use standard library |
| 23:27.29 | MDK | :> |
| 23:27.45 | MDK | joking |
| 23:28.13 | MDK | jazelle looks like a fake sliver bullet |
| 23:28.17 | xorAxAx | did you know that brainfuck on .net is faster than compiled to C? |
| 23:28.22 | xorAxAx | i.e. C# vs. C |
| 23:28.36 | xorAxAx | (it was a simple prime sieve that was benchmarked) |
| 23:28.57 | MDK | yeah, that's not surprising |
| 23:28.58 | xorAxAx | and the brainfuck code was compiled in both cases |
| 23:29.03 | xorAxAx | oh, really? |
| 23:29.09 | MDK | I did recently some tests with oo-oriented code |
| 23:29.14 | xorAxAx | the JIT seems to be really smart |
| 23:29.26 | MDK | I tested C#, GObject and my own home-coocked object oriented system |
| 23:29.27 | xorAxAx | brainfuck is not ooo :) |
| 23:30.06 | MDK | hmm |
| 23:30.17 | xorAxAx | you know, less than 10 commands :) |
| 23:30.40 | MDK | well, you can always create a 10-line program to prove absolutely anything |
| 23:30.47 | MDK | like Ruby being faster than C |
| 23:30.51 | xorAxAx | it was a prime sieve |
| 23:31.01 | xorAxAx | no, we are not comparing languages |
| 23:31.08 | xorAxAx | we are comparing machines |
| 23:31.19 | xorAxAx | the x86 and the CLR |
| 23:31.22 | MDK | yeah, I'm not arguing really |
| 23:31.26 | MDK | I believe that's true |
| 23:31.40 | MDK | but the point is, that in "real life" many other factors come into play |
| 23:31.41 | xorAxAx | i was surprised ... bf code is 99% pointer deferencing |
| 23:31.46 | xorAxAx | of course :) |
| 23:32.05 | MDK | like the whole java library stack, display stack, etc. |
| 23:33.20 | xorAxAx | "There is something about Helsinki that reminds me of Twin Peaks, but overally it's a great place to live. " haha |
| 23:33.25 | MDK | we're working now on bringing mono + gtk-sharp to 770 |
| 23:33.31 | xorAxAx | nice |
| 23:33.40 | xorAxAx | do you know about a successor of the 770? |
| 23:33.51 | xorAxAx | with a faster cpu for example :) |
| 23:33.58 | MDK | patience my friend |
| 23:34.29 | xorAxAx | yeah, many 770 users seem to miss patience - "oh, my linux app is too slow", "when will i be able to buy the new device" |
| 23:35.12 | *** part/#openmoko tapo (n=tklopp@dslb-084-060-054-020.pools.arcor-ip.net) |
| 23:35.53 | MDK | hardware is hardware, software matters ;) |
| 23:36.43 | xorAxAx | oh, you are working on insane fast gcc optimisations for the arm platform? ;-p |
| 23:36.54 | MDK | no no, not at all |
| 23:36.58 | xorAxAx | hehe :) |
| 23:37.06 | MDK | I work with high-level desktop stuff |
| 23:37.25 | MDK | and I see HUNDREDS of places where we could make it faster and better |
| 23:38.36 | MDK | you know, people used to do magic with 64kb of ram... ;) |
| 23:39.56 | xorAxAx | yeah, the 770 is a perfect target for the demo scene |
| 23:40.01 | xorAxAx | but they still seem to prefer windows |
| 23:40.17 | xorAxAx | for 64kb |
| 23:40.47 | jaebird | MDK: of diva fame? |
| 23:41.00 | MDK | well, I guess arm has different set of tricks |
| 23:41.14 | MDK | I mean, those guys know a shit load about i386 |
| 23:41.23 | xorAxAx | sure |
| 23:41.34 | MDK | but I guess arm is a bit different |
| 23:41.39 | MDK | jaebird: true |
| 23:41.42 | jaebird | MDK: project still alive? |
| 23:42.28 | MDK | jaebird: a bit stalled |
| 23:42.35 | MDK | jaebird: i burn-out a little |
| 23:42.42 | jaebird | :) don't we all |
| 23:43.01 | MDK | jaebird: I haven't received any contributions, was hacking all this alone, this kind of stuff |
| 23:43.17 | MDK | and eventually realized gstreamer is bad |
| 23:43.18 | jaebird | still hoping it will materialize! |
| 23:43.22 | MDK | for this kind of things |
| 23:43.39 | jaebird | not sure where gstreamer is going actually |
| 23:44.04 | jaebird | i was hoping for a cross-platform multimedia framework |
| 23:44.17 | jaebird | nobody seems to really care all that much |
| 23:44.37 | MDK | well, it's all about design I guess |
| 23:44.46 | MDK | I know those guys, some of them personally |
| 23:44.51 | MDK | they're great hackers and everything |
| 23:45.01 | MDK | but they just completely missed the design |
| 23:45.09 | MDK | it's a very bad design |
| 23:45.26 | jaebird | i know about bad designs :) |
| 23:45.38 | jaebird | been there, done that |
| 23:46.03 | MDK | yeah, design matters |
| 23:46.06 | jaebird | so without gstreamer, what is diva going to use? |
| 23:47.23 | MDK | I was hacking on my own small engine |
| 23:47.38 | MDK | but that's a long story |
| 23:48.08 | jaebird | ok...well didn't mean to fork this...i'm hangin out on #diva anyways! |
| 23:48.30 | MDK | ;) |
| 23:48.56 | MDK | maybe one day we'll assemble a dream team and get this done right |
| 23:49.06 | jaebird | totally |
| 23:49.41 | jaebird | until then I get to play with my E2 |
| 23:49.57 | MDK | E2? |
| 23:50.04 | jaebird | Moto E2 Rokr |
| 23:50.13 | jaebird | one of the _linux_ phones |
| 23:50.31 | jaebird | my plan is to dump OpenMoko on it once it is ported |
| 23:50.31 | MDK | ;) |
| 23:50.45 | jaebird | right now it is my ePod |
| 23:51.14 | jaebird | but does not yet support ogg...even tho they use the helix framework |
| 23:51.23 | jaebird | bazaar |
| 23:51.47 | MDK | yeah, the ogg support is kinda strange thing |
| 23:51.51 | MDK | nobody does it |
| 23:52.11 | MDK | I think vorbis is a nasty patent shit-hole |
| 23:52.17 | MDK | and nobody really wants to touch it |
| 23:52.18 | jaebird | maybe suits are scared of the whole patent thing |
| 23:52.26 | jaebird | diddo |
| 23:52.45 | MDK | that's really anoing |
| 23:53.06 | MDK | true inovation is being killed |
| 23:53.11 | jaebird | yeah, patents really torc me off |
| 23:53.12 | xorAxAx | MDK: do you know directshow? |
| 23:53.28 | MDK | xorAxAx: yes, to some extent |
| 23:53.42 | xorAxAx | would it be better than gstreamer in your opinion? :=) |
| 23:53.46 | xorAxAx | (the design) |
| 23:54.08 | jaebird | COM...yuck |
| 23:54.14 | xorAxAx | yeah, COM, cls ids ... |
| 23:54.14 | MDK | paradoxally, I think that yes |
| 23:54.25 | MDK | but that's like choosing between lesser evil |
| 23:54.25 | xorAxAx | but thats just an implementation detail somehow |
| 23:54.36 | xorAxAx | you could design that away :) |
| 23:54.57 | xorAxAx | so somebody should start gshow :) |
| 23:55.06 | MDK | the problem that people seem not to get, is that you CAN'T create a framework that is a good playback/show framework and a good editing framework at the same time |
| 23:55.07 | jaebird | why didn't the gstreamer folks use directshow as a design model? |
| 23:55.15 | MDK | it's just two separate things really |
| 23:55.44 | MDK | gstreamer is a player framework. Possibly a transcoding framework. But using it for editing is a pain. |
| 23:56.05 | xorAxAx | hmm, does any video editing system use direct show internally completly? |
| 23:56.17 | jaebird | if you went for an editing framework first...i bet a show framework would evolve |
| 23:56.22 | xorAxAx | i mean any larger one |
| 23:56.22 | MDK | jaebird: they used. Ithink Ronald Bultje had something to do with directshow |
| 23:56.36 | MDK | but ronald is no longer actively designing gstreamer |
| 23:56.39 | xorAxAx | the company i am working for at least wrote one that does, but its a very simple one |
| 23:57.44 | xorAxAx | i just remember that premiere felt completly independant from directshow |
| 23:57.46 | xorAxAx | when i used it |
| 23:58.02 | jaebird | is the tricky part the render pipeline? |
| 23:58.22 | jaebird | i wrote a source capture filter once |
| 23:58.34 | MDK | well, it's th e |
| 23:58.44 | MDK | *the real-time playback pipeline |
| 23:58.58 | xorAxAx | jaebird: with directshow, an app simply says "i have a directshow object here, please connect it to the directsound etc. devices automatically" |
| 23:59.10 | xorAxAx | and then it instantiates the filters and connects them |
| 23:59.11 | jaebird | seems the clock is problematic for rendering |
| 23:59.15 | MDK | yeah, gst does that too |
| 23:59.33 | MDK | the problem is: "hey, the cut is near. I want to disconnect the object, and connect a new one" |
| 23:59.46 | xorAxAx | that working nicely in dshow as well |
| 23:59.54 | xorAxAx | *except if you see any segfaults2H |
| 23:59.59 | MDK | and than, in gst, you're in dark pit of threads, locking, async events, etc. |