00:00:00 --- log: started forth/03.11.20 03:04:29 --- join: XeF4 (~xef4@eukaryote198.gprs.suomen2g.fi) joined #forth 03:44:25 --- quit: XeF4 ("pois") 05:08:44 --- join: schihei (~schihei@p5085DAFC.dip.t-dialin.net) joined #forth 05:15:41 --- join: Serg_Penguin (~z@212.34.52.140) joined #forth 05:20:00 quack uack 05:20:01 :) 05:35:45 --- quit: I440r ("going to work -->") 05:43:54 --- join: elmex (~elmex@elmex.x-paste.de) joined #forth 05:43:58 helo?! 05:47:33 --- part: z0d left #forth 05:47:43 hmm 05:47:49 hel lo 05:48:19 the forth FAQ @ forth.org isn't avaible.. 05:55:11 hm, i like to program networking programs... is it possible to interface sockets somehow? is there any standart facility which makes it possible? or is forth for other purposes? 06:11:26 dup dup dup di di dup 06:13:58 --- join: I440r (~mark4@saturn.vcsd.com) joined #forth 06:16:59 elmex: yes, in RU spforth 06:17:05 www.forth.org.ru 06:17:38 i attempted to write a trivial news client, but failed th grok it's socks i-face 06:18:57 hm, i somehow would like to use gforth 06:19:56 elmex, why gforth specifically ? 06:20:06 and whats stopping u ? 06:20:13 simple reason: because i like gnu software 06:20:19 lol 06:20:24 isforth.clss.net :) 06:21:44 --- quit: Robert (Read error: 60 (Operation timed out)) 06:24:15 --- join: Robert (~snofs@c-255a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 06:26:23 --- join: madgarden (~bughead@216.94.153.178) joined #forth 06:27:50 hm... i guess..forth is way too lowlevel for my taste :) (i'm more a script programmer, and like C... ;) 06:31:17 --- quit: schihei (Client Quit) 06:31:39 --- join: schihei (~schihei@p5085D976.dip.t-dialin.net) joined #forth 06:33:57 bbl 06:33:58 --- part: elmex left #forth 07:01:54 --- quit: Serg_Penguin () 07:13:40 --- quit: schihei (Client Quit) 07:14:22 --- join: schihei (~schihei@p5085D7E7.dip.t-dialin.net) joined #forth 07:22:59 --- join: melinda (melinda@melinda.usercloak.freenode) joined #forth 07:48:09 --- quit: schihei (Client Quit) 07:49:30 --- join: schihei (~schihei@pD954894A.dip.t-dialin.net) joined #forth 08:15:19 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 08:40:20 --- quit: I440r ("Leaving") 08:41:52 --- join: I440r (~mark4@saturn.vcsd.com) joined #forth 08:49:42 --- quit: I440r ("Leaving") 09:26:29 --- join: kc5tja (~kc5tja@66-91-231-74.san.rr.com) joined #forth 09:26:29 --- mode: ChanServ set +o kc5tja 09:27:13 Well, I've decided to start writing my own humane text editor for Linux. (in C, using GTK as its toolkit, and arch as the source code management system) 09:28:00 Since it's written in C, not much else will be said about it here, except to let people know that's my current project. :) 09:28:59 why not python? 09:30:05 I'm suprised you're starting another big project 09:31:21 This is a project I've been wanting to work on for a while now, if you'll recall. 09:31:43 * kc5tja has always wanted a humane editor, and so now I'm going to write one. 09:32:16 Why not Python? Because I feel C will offer smaller, faster software. 09:32:28 I think it's great you're making one 09:32:35 In fact, speed is pretty critical issue for me. 09:32:41 s/is crit/is a crit/ 09:32:59 Plus, it helps keep my XP programming skills up to date. 09:33:21 I have it opening a window, with nothing inside it. :) And it's a ROX AppDir already. :) 09:33:22 kc5tja: à la THE? 09:33:31 is there some incompatibility between python and XP? 09:33:32 ROX? 09:33:43 ianP: Inspired by THE, yes, but not as far reaching. I'm aiming first and foremost for text editing applications, not a whole environment. 09:33:52 kc5tja: so more practical ;) 09:33:56 Though I do intend on extending it to include a complete environment at some point in the future. 09:34:08 Yes, I'm going to add features to it only as I need it. 09:34:10 kc5tja: nice :) 09:34:14 And I've decided on REXX as its scripting language. 09:34:16 kc5tja i suppose there's not much to the far-reaching-ness of THE though, besides incompleteness. I wish he'd freaking port to OSX already 09:34:29 it was fun to play with when i had os9 running 09:34:31 ianP: re ROX; http://rox.sourceforge.net 09:34:33 heheh 09:34:44 ianP: play with what? Rexx? 09:35:03 I've tried THE on Mac os 9, and it crashed so much it wasn't fun to play with 09:35:05 Oh, OS9 as in MacOS 9. Sorry. 09:35:27 Well, here's the deal. 09:35:32 Stability is the ultimate requirement. 09:35:35 Speed is the #2 requirement. 09:35:40 it didnt crash 09:35:42 form e 09:35:46 kc5tja: what do you need the speed of C for? searching? 09:35:53 --- quit: haroldo_ (Read error: 104 (Connection reset by peer)) 09:36:12 seems to me a text editor is something that mostly sits there and waits for user input 09:36:20 Herkamire: Among other things. The humane concept presents a "flat" file concept to the user, and if you have hundreds of documents in the humane environmnent, your session is going to be *huge*. 09:36:41 Herkamire: Also speed in rendering the display too. 09:37:00 have you considered SDL for graphics? 09:37:08 and input 09:37:10 Also convenience in accessing the document memory. Python must go through string or sequence accessors to do this, which just slows things down. 09:37:34 Herkamire: Yes, but unless using GTK becomes overbearingly hard, I don't see a need to change from it. 09:38:08 oh yeah. I tried to write something in python where I needed direct access to memory, sometimes bytes sometimes words, and sometimes half-words. it was a major pain. I translated it into C :) 09:38:20 My goal is to make 'qm' (what I'm calling it; short for QuasiMode) as fully ROX integrable as possible. 09:38:43 I already have the AppDir finished. 09:38:49 It even compiles itself the first time. :) 09:39:02 And I compile with -Wall, and it's a development *requirement* that *ANY* warning *MUST* be fixed before release. 09:39:18 can you do the keyboard handling you need for THE with gtk? 09:39:48 If I write my own widget, I'm positive I can. GdkEvents mirror all the raw X11 events closely. 09:39:57 kc5tja - nice 09:40:04 cool 09:40:39 Once I get this thing even *minimally* functional, I'm planning on opening it up for Bazaar-style development. 09:41:07 word. 09:41:18 i like the way you worded that :) 09:41:23 hmmm.... I still think you should use SDL. I haven't done enough with gtk to really know, but SDL is so simple. 09:41:30 only problem is you'll have to do your own text rendering 09:41:40 SDL also requires me to work up my own complete set of widgets, and font engines. 09:41:49 Right. 09:42:01 It may still come to that, I don't know. 09:42:03 what widgets? 09:42:08 will it be easy to port to another gui framework? 09:42:19 ianP: That depends on how friendly GTK is towards that concept. 09:42:45 i suppose portability often depends on code quality too 09:42:54 with good design much more things are possible i find 09:43:02 --- join: haroldo (~haroldo@r200-40-218-202.adsl.anteldata.net.uy) joined #forth 09:43:03 Being that it is developed with XP principles, while still meeting most of the GNU coding standards, GUI code will necessarily be separated from the core logic of the code. 09:43:34 The overall architecture of the system will be as MVC as possible. 09:43:52 right on 09:44:32 Herkamire: If I write using SDL instead of GTK, then it wouldn't support resizable windows, I don't think. Resizable windows are a must for productivity applications, even though I personally use Ion. 09:44:57 SDL can be resizable 09:45:08 Also, ? 09:45:12 err 09:45:13 ? 09:45:15 I didn't know that. 09:45:19 my koch fractals thing is SDL in a resizable window 09:45:27 hmm 09:45:41 How does SDL communicate that a window has been resized? 09:46:20 (it would be kind of neat to be able to use qm in full-screen mode, I suppose, as when working exclusively with console applications...) 09:46:22 http://jason.herkamire.com:5000/svn/koch/sdl.c 09:47:07 I think if you used SDL you could have your editor work on the framebuffer of many machines (SDL can work through SVGALIB right?) 09:47:41 SDL_SetVideoMode(w, h, 0, SDL_RESIZABLE) 09:48:18 Herkamire: I don't think so, but I know it can use fbcon and GGI directly. 09:48:29 What is the deal with SDL_LockScreen()? 09:48:41 don't know. I just followed docs. 09:48:59 there's a tarball for that project here if you want it: http://herkamire.com/jason/downloads/kochfractals-0.2.1.tar.bz2 09:49:05 And can I get SDL to deal with framebuffer conversions for me? That is, I don't want to have to deal with 8-vs-16-vs-24big-vs-24little-vs-32 bits per pixel modes. 09:49:09 or feel free to browse http://jason.herkamire.com:5000/svn/koch/ 09:49:22 fwoom! 09:49:40 fwoom? 09:49:49 what do you mean conversion? you mean change the bit depth of the screen? 09:51:11 No, I mean that I don't want to have to deal with 8/16/24b/24l/32 bits per pixel. 09:51:24 I don't want to have to have 5 different back-end renderers. 09:51:47 you can ask SDL to change the bit depth http://sdldoc.csn.ul.ie/sdlsetvideomode.php 09:51:49 kc5tja: yeah, rendering one back should be plenty ... 09:52:02 but you won't always get it. 09:52:10 Herkamire: I don't think you're understanding me. 09:52:13 with SDL you usually deal with the different pixel formats 09:52:35 I do not want to have five different back-end renderers in my program. :) I want *one* function that pokes pixels into memory, and that's it. 09:52:58 for that you need to have memory in a standard format. 09:53:01 I understand that having five offers the fastest visual performance, but I'm not writing a game either. :) 09:53:18 I don't think you can have that. 09:53:27 Herkamire: Maybe with a SWSURFACE instead of a HWSURFACE? 09:53:35 you could write one backend and five blitters to copy what the backend makes onto the display 09:53:48 Herkamire: That's the same as having five back-end renderers. 09:53:50 blitter? 09:53:50 SDL may have those blitters 09:54:24 MysticOne: A blitter is a piece of software or hardware, designed to block-move bits from one bitmap to another, performing various transformations in the process (e.g., changing colors, handling transparencies, etc) 09:54:36 kc5tja: that might work. if you ask for swsurface, SDL may translate for you. then you can write to memory in whatever format you want. 09:54:53 I've forgotten a lot about SDL internals. 09:55:02 I'm sure they've probably updated the API a bit too. 09:55:12 then the only think you'd have to bother with is setting the 8-bit pallete. 09:55:22 16-bit. 09:55:30 interesting 09:55:31 I need at least 16 bits to handle transparent text properly. 09:55:45 (I need control over the RGB values of each pixel precisely) 09:56:28 Is there any advantage to using an 8-bit screen instead of a 16-bit screen? Other than reduction in consumed memory? 09:56:49 If so, I can use my pseudo-512-color screen technique to fake a 3-bit-per-channel display mode. 09:57:44 Otherwise, I'd rather just use 16 bits per pixel, to make the math easy on the computer. 09:58:06 kc5tja: don't really know. I figure it's cool to support 8-bit display if you can because that's what some people have. I don't think many people use it though. I think gamers sometimes switch the monitor to 8 bit because blits are faster (lower bandwidth to gfx card) 09:58:41 kc5tja: if you need that many colors, than you can just require at least 16 colors. 09:58:45 Most gamers I personally know have 32-bit displays, and rely exclusively on the card's GPU to do all the work needed. :) 09:58:57 linux terminal is usually in 8-bit 09:59:08 Herkamire: You mean fbcon? 09:59:21 I've forgotten what fbcon is again 09:59:31 Frame buffer console 09:59:37 yeah. that 09:59:44 It's what your Linux uses because YOUR computer lacks a proper text-mode. 09:59:50 (not that that's a bad thing...) 10:00:18 it has a text mode ;) you get it by holding down apple-option-O-F 10:00:49 That's not text mode. 10:00:53 That's OpenFirmware. 10:01:01 The card is still sitting in a graphics mode to display it though. 10:01:31 I think the text renderer for that OF text is in the graphics card. not sure though. 10:01:55 doesn't matter. still probably not a text mode. 10:02:21 At any rate, point taken regarding video mode selection. 10:02:48 yeah. I didn't even think about that (that most machines are in text mode not 8-bit) 10:03:17 so svga has to change the graphics mode 10:03:25 Yep. 10:03:39 That's why PCs have to go ***CLACK!!*** when the monitor changes video frequencies and whatnot. :) 10:04:12 Those are a bunch of relays switching the CRT's horizontal and vertical refresh rates to keep up (and so as to not destroy the monitor's flyback circuitry in the process). 10:05:04 CRTs are annoying 10:05:25 But, I'm kind of wondering about the whole 8bpp thing though, especially if qm is going to work also for console operation. 10:05:26 although it is cool to be able to go low-res for a game to run smoothe 10:05:51 You mean, without having to wear magnifying glasses to play it. :) 10:06:30 Nahh, I'll just choose 16-bpp mode stuff now, and just let the open source developers hack it to their desires. 10:06:32 :) 10:06:36 Make it easy on myself. 10:06:45 But SDL is looking a bit more attractive to me now. 10:07:08 My only concern is fonts. 10:07:25 (and rendering the text on the screen) 10:07:26 do you want multiple fonts? 10:07:41 do you want variable width font(s)? 10:07:44 At some point, I'm going to have to, I think. 10:07:51 Maybe not from the get-go, but in the future. 10:08:10 Also Unicode support. 10:08:19 really 10:08:34 wow. I wouldn't bother with any of that. 10:08:49 Well, this is what I'm saying: *I* am not, to start with. 10:08:50 you sound kinda serious about this ;) 10:09:00 But to be a truely useful editor, I think it needs to be supported at *some* point. 10:09:19 I mean, even VIM supports Unicode now, IIRC. :) 10:09:37 I wonder if there is any advanced text rendering stuff for SDL 10:09:49 vim does? :) 10:10:08 Well, one can use the FreeType library to render TrueType fonts, and I know that supports Unicode character sets if the font does. 10:10:27 But how it handles right-to-left and other nationalized text rendering issues, I don't know. 10:11:14 BTW, I notice you're using SVN. :) 10:11:18 * kc5tja is using arch. 10:11:26 Down with CVS!! 10:11:43 CVS haters of the world -- COMMIT! ;P 10:12:52 yeah. svn has been much better to me that cvs 10:13:06 svn looks like they are nearing beta finally :) 10:13:25 vim does indeed support Unicode 10:14:01 According to savannah.gnu.org/projects/gnu-arch, arch is actually production ready. I have personally found only one bug with it, and it's annoying, but not project killing. 10:14:33 arch is the one that's written in bash, and has weird ALL-CAPS directories in your working copies right? 10:15:56 I've only found one bug with svn (that's easily avoided) and I'm using a version that is very old (there have been about 20 bugfix releases since) 10:16:52 arch has been rewritten in C, and while it has distinct naming conventions (e.g., files beginning with + are "precious, non-source" files, while files starting with ',' are junk files, etc), nothing in my directory trees are all-caps. 10:17:06 I do not believe the Bash version is being maintained anymore. 10:19:37 cool 10:21:04 arch looks very cool 10:22:58 there's sdl_ttf which renders truetype fonds in SDL. not sure if it does unicode though. http://www.libsdl.org/projects/SDL_ttf/ 10:27:33 looks like it does unicode 10:41:40 If it uses FreeType, it should. 10:41:48 But I think rendering characters takes more than just producing the graphics. 10:41:54 So, for now, what I'll do is this: 10:42:07 when importing text files, I'll expand each character to a 16-bit or 32-bit word. 10:42:52 BUT, I'll only support proper ASCII codes for the time being. Anything above that is rendered according to the font used, but is not in itself "supported officially" by the program. 10:42:55 That'll be a to-do item. 10:43:31 In retrospect, I think I have one more compelling reason to support SDL instead of GTK: it's vastly more cross-platform. 10:43:35 SDL works on AmigaOS too. :) 10:43:47 (and, IIRC, OS/2, BeOS, et. al.) 10:43:56 So it greatly improves the portability of qm. 10:44:04 I think that has convinced me. 10:44:30 Oh, and native MacOS too. 10:49:05 OXS native? 10:51:29 OXS? 10:51:41 You mean OS X? I would assume SDL has been ported to OS X. 10:54:00 --- join: Speuler (~Speuler@node-d-9180.a2000.nl) joined #forth 10:54:17 g'day 10:56:52 re 11:00:32 hi 11:02:24 Okay, the final architecture won't be true MVC; it'll be Document/View instead (basically, the View includes the controller built-in). 11:02:41 I can't think of a way to split the controller functionality out of the View proper. 11:03:14 If I come up with something, I'll use it. 11:16:54 clear 11:16:56 gahh! 11:17:58 Crap. I haven't even gotten started yet, and I'm already finding myself in need of a good object system. :/ 11:18:24 --- quit: schihei (Client Quit) 11:38:31 anyone here know of any embedded systems (x86 based) that can be fanless and get about 60-70mbps of network throughput? :) 11:40:41 Soekris? :) 11:40:44 I'm guessing? 11:42:00 no 11:42:06 not fast enough :( 11:42:10 maxes out at about 20-25mbps 11:44:57 In that case, then I guess you're out of luck. :/ 11:45:10 BTW, are you sure you mean mbps and not Mbps or MBps? 11:45:12 :) 11:45:22 millibits per second are pretty slow throughputs. :) 11:45:43 megabits :) 11:45:51 there's mini-ITX ... 11:45:57 but damn, I don't want fscking sound on it! or any of that crab 11:53:38 Well, why is Soekris only getting 25Mbps throughput? 11:53:42 What CPU and what clock-speed? 11:53:57 AMD ELan 133 MHz (roughly a P120) 11:58:46 kc5tja: spaced out for a while there. that's great you're going to use SDL :) 12:00:06 Herkamire: Well, I'm already running into issues with respect to objects in the system. 12:00:23 lol 12:00:25 * kc5tja has to basically reinvent an object system, or switch to C++, to achieve my goals. 12:00:32 python then? 12:00:37 No. 12:00:43 Python is too inconvenient for this. 12:01:19 for some oo-likeness, you might consider using quans 12:01:46 those are words, consisting of a list of xts 12:02:12 Speuler: I'm not writing this in Forth. It's written in C. 12:02:13 could view them as objects 12:02:40 * Speuler checks wether he has joined the wrong channel 12:03:04 hehe 12:03:07 kc5tja, try #c 12:03:29 Speuler: There is no need. 12:03:31 It's not like people talk much Forth in here anyway. :P 12:03:35 Notice that I'm not asking for assistance. 12:03:47 I'm merely reporting thoughts. 12:04:02 I've made object systems in C before, and I can do it again. 12:04:13 I just need to think things through a bit. 12:04:17 i made those in forth :) 12:04:34 So have I. 12:04:53 thus, no problem 12:06:09 --- join: thin (thin@bespin.org) joined #forth 12:06:17 kc5tja: i emailed you at falvosa yahoo.. 12:06:25 have you gotten that email yet? 12:06:52 thin: I no longer check that account. Use kc5tja at arrl dot net instead. 12:07:25 Speuler: The 'problem' is that when writing for GTK, I had an object model already usable -- GTK's. Now I don't. So now I have to make one. 12:07:51 Since I had not counted on doing that today, it kind of screws up my list of things to do. 12:08:45 i'd figure so does explaining to us what you need to do :) 12:09:03 kc5tja: how long do you think it will take you to make this editor? (the basic features anyways) 12:09:16 p.s. email sent 12:11:13 Speuler: The object model is to support a model/view/controller triad, for the most part. It is also required to support object prevalence, which allows the editor to come as close to being orthogonally persistent as one can get under Linux. 12:11:17 thin: I have no idea. 12:11:45 Speuler: In the future, I'd like to CORBA-enable the editor, so as to allow remote scriptability. 12:15:07 thin: I'm working on it in my spare time, as you might expect. 12:15:10 i've been thinking, i don't want to know the same amount of computer/software trivia that you guys know in order to accomplish the same amount that you guys can.. it seems quite ridiculous to have to learn mutliple interfaces to multiple software packages when all i want is one complete interface (i.e. forth or something else) 12:15:37 i'm talking about "the dream" here :P 12:16:18 kc5tja: whats happening with some of your other projects? Doesn't seem like you have much free time since you have quite a few projects.. 12:16:31 Well, the Singularity isn't too far off. 12:17:31 kc5tja: maybe you should make a throwaway THEish, rather than worrying about future features like unicode, etc.. 12:17:34 thin: qm is a project designed to accomplish several goals related to Dolphin. 12:17:57 First, it's an experimental editor with humane concepts in mind, designed to help explore the construction and manipulation of a flat, file-less user interface. 12:18:11 It is Dolphin's primary goal to be as user friendly as possible, and I feel this is useful in and of itself. 12:18:14 madgarden: i don't know what Singularity is, and i don't think i want to :) 12:18:32 Hee. :D 12:18:45 Second, the object model I was going to adopt for it (now that I'm using SDL) ties right into my ideas for a very, very high performance CORBA wire-level compatible object model. 12:18:46 * thin expects its some religious/programming combo 12:19:15 thin: I too wish I didn't have to learn all these stupid interfaces. that's why I'm making an OS. Ironically to do so I have to learn all sorts of hardware interfaces 12:19:27 herkamire: :) 12:19:32 This ORB implementation is itself Dolphin related, as I have plans of using CORBA extensively throughout Dolphin. 12:20:48 CORBA sounds like a pretty mainstream thingie and i'm usually suspicious of mainstream thingies because they're usually bloated. is CORBA simple & rational? 12:20:59 --- quit: warpzero (Read error: 110 (Connection timed out)) 12:22:47 thin: Depends on how simple and rational you want it to be. 12:23:06 CORBA is a tool. 12:23:10 It does its job. 12:23:13 It does it well. 12:23:27 But like anything that is a standard, it has to have implementations. 12:23:39 And implementations can vary from vendor to vendor. 12:23:59 MICO, for example, gave the kDE developers the reputation that CORBA is slow, resource intensive, and utterly unfriendly to programmers. 12:24:21 ORBit, however, is the core of Gnome's CORBA infrastructure, and is extremely simple to write software for (well, simple within the context of CORBA). 12:24:27 Like all standards, there is a learning curve. 12:24:30 That's just the way things are. 12:24:51 Anyway, ORBit is at least an order of magnitude faster than MICO, if not faster. 12:25:04 It's much lighterweight too, and supports only a proper subset of the whole CORBA standard. 12:25:33 How does that compare to XMLRPC? 12:25:35 What I'm looking to do is even lighter-weight still; I'm looking to support ONLY the wire-level protocols and basic capabilities of CORBA-compatible objects. 12:26:16 madgarden: It doesn't. CORBA is so overwhelmingly superior to XML RPC that they're not even considered on the same level. 12:26:34 Well, there you go! ;) 12:26:37 would chuck moore use CORBA or would he invent something more forthlke? 12:26:54 how come you are going to use REXX for the scripting language of the editor and not forth? 12:27:15 thin: Approachability by newbies and experts alike. 12:27:45 thin: I have considered using Guile, since qm is going to be GNU licensed, but Guile, being a Scheme dialect, is going to thoroughly confuse people. Utterly. 12:27:58 (Guile seems to be the perfect name for it then...) 12:28:00 Forth is cool, but something REXX or BASIC is *fun* to use. 12:28:27 Also, Forth lacks innate string and memory management features which are a requirement for text editing. 12:28:42 Forth also lacks a system by which it can 'address' other programs to control them with. 12:29:15 madgarden: depends on what implementation you've used.. MUF (forth-like) is more fun than BASIC, altho i have no experience with REXX 12:29:16 REXX can/will provide qm with a scripting language at least as powerful as elisp is for Emacs. 12:30:46 REXX is not a fast language. But it IS a good language for string manipulation. (In fact, it even internally represents numbers as strings, which gives it a unique advantage over other languages right out of the box: infinite precision, and it can accurately represent 1/3 and 1/7 without round-off errors!) 12:30:47 thin: And well, my Forth-like (forthy) is fun to use, and has strings and so forth as well, I imagine it's a lot like MUF. But (ANS) Forth itself wouldn't be a good choice, I'd think. Now, a high-level Forth extension, sure. 12:31:21 And string manipulation is precisely what you want for text editing purposes. 12:31:26 the funnest forth implementation to date was probably the one that the engineers of iTv were trained on 12:31:31 machienforth 12:31:39 There's nothing funny about it. 12:31:44 i didn't say funny 12:31:50 I find it extremely elegant. 12:31:54 Oh, sorry. I misread it. 12:32:03 * kc5tja is bleary-eyed right now. Didn't get much sleep. 12:32:10 If I seem crabby, that's why. 12:32:18 Please don't take it personally. 12:32:20 yeah.. its a ton of fun because its extremely elgant, simple, no bullshit, no noise, no worries about esoteric interfaces 12:32:25 its just the man and the machine 12:33:28 i won't take it personally :P 12:33:28 * kc5tja thinks of abandoning the classical inheritance type object system in favor of aggregation-based system. 12:33:48 Aggregation/composition is the shiznit. 12:34:00 And, it's perfect for a C implementation. 12:34:00 * kc5tja is shocked to hear someone else in favor of aggregation. :) 12:34:13 Well, I don't much like C++, so static inheritance annoys me. ;) 12:34:21 (especially if I have to do it in C) 12:34:26 * kc5tja nods 12:34:29 isn't it well known that inheritance is a lousy system? 12:34:30 :P 12:34:50 that was rhetorical 12:34:51 I like the idea of determining the type of an object based on what interfaces it actually supports, versus what it claims to support. 12:35:18 thin: It's actually a good system, but when multiple-inheritance comes into play, it can cause major headaches. :) 12:35:45 It also sucks when you only want to inherit parts of something. 12:36:39 I guess that if I'm going to support an aggregation-based object system, I would probably end up with a COM-clone. 12:36:48 COM is pretty damn simple in that respect. 12:37:07 COM == Component Object Model 12:37:21 You going to write your own registry too? :) Love those GUIDs! 12:37:41 Why do that? The Unix filesystem is as good a registry as anyone will ever need. 12:38:26 http://sourceforge.net/projects/gcom 12:38:34 (I'm surprised they're still hosting it!) 12:39:10 * thin is eagerly waiting for a reply to the email he sent kc5tja ;) 12:39:23 if you're too tired to respond today that's fine.. 12:40:00 I would be happy to reply if the POP connection limit was not so easily reached on my mail server. 12:40:05 It's been that way for HOURS now. 12:40:47 Yea, I don't much like the registry. ;) 12:41:26 Well, a registry is required. 12:41:43 Otherwise you couldn't say, "Give me an object with a FooTextWidget interface." 12:42:07 There could be any number of implementations of FooTextWidget interfaces; the registry needs to maintain that linkage. 12:42:19 What sucks is that MS decided to use the registry for **EVERYTHING** else too. 12:42:38 MS could just as well have used the filesystem. 12:43:07 It's THE registry I'm opposed to. Sure, throw all my eggs in one basket, why not! Then they're all in one place for me to drop... er... 12:43:09 Yep. 12:43:30 Gosh, that's only version 0.2 up there. 12:43:37 All these years, I had 0.3, but I apparently forgot to upload it. 12:44:00 Just one more thing to throw into the arch repository. 12:44:01 :) 12:49:21 extreme programming seems to be out. at least, interest in it seems to be close to 0 12:49:57 i conclude this from forth wiki page popularity 12:50:18 Speuler: I fail to see how that relates? 12:50:25 99% of Forth coders do not practice XP anyway. 12:50:28 the was exactly one hit in 3.5 months on the xp page 12:50:50 Try http://www.xpsd.com -- that's a whole Wiki devoted to XP development in the San Diego area. 12:51:04 It's not as busy as it used to be either, but the XPSD meetings are as large as they've ever been. 12:51:35 i translate this to "popularity didn't increase" 12:52:29 Depends on your point of view. 12:52:47 t'is just a an observation 12:52:48 I don't think you can make that judgement, even with the evidence you have on hand. 12:52:59 Any more than I can say that it has increased. 12:53:19 But here's what I do know -- bookstore shelves are packed with XP-related books. So I know they're selling. 12:53:32 does chuck moore practice XP? :P 12:54:26 thin: His practices are largely in leage with XP's practices. But he does not practice XP itself, no. 12:54:42 XP is a person-to-person development methodology; no cubicle walls in offices, programmers talk directly to customers, etc. 12:54:43 i like the "there must be food" bit 12:54:49 :) 12:54:53 Chuck prefers solitude in his work. 12:56:45 speuler: yeah, you're the food junkie of the channel ;) 12:57:05 and always candies at meetings 12:59:29 kc5tja so you actually got the email or you havent? 12:59:44 No, not yet. 12:59:55 I can't check any of my mail, because, "POP3 connection limit exceed." 13:00:21 I've never seen my mail server this bad before. 13:03:04 kc5tja: ok, read the email at http://thin.bespin.org/idea.txt 13:05:52 So I guess that each person on the site would have their own blog, and the front page just filters the blogs for the most recent posts by each user? 13:05:58 (sorting them in chronological order, I guess) 13:09:56 I think the forth community is too small for a system that formal. I think a wiki would be good 13:11:15 Oooh, I would love a Forth wiki. 13:11:16 I am happy to discuss my ideas and experiments on forth, but I am not going to write regular articcles about it or even bother signing up as an editor on a site. 13:11:53 madgarden, you're aware there are forth wikis ? 13:12:36 what's our wiki status? I've seen at least one, but I'm not sure if it was going to stick around 13:12:56 Yes,I have seen a couple Forth wikis out there, though I seem to lose track of them. They didn't seem too active. 13:14:45 --- join: warp0x00 (~warpzero@dsl.142.mt.onewest.net) joined #forth 13:16:20 --- join: I440r (~mark4@saturn.vcsd.com) joined #forth 13:19:05 we could pick a wiki and promote it with this site 13:19:14 s/this site/this chatroom/ 13:20:43 hgn 13:27:49 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 13:29:52 --- quit: tathi (Client Quit) 13:29:54 How about an existing general software development wiki? 13:32:54 I think the best wiki stuff I saw for forth was on some general wiki. possibly the original wiki 13:35:09 c2.com 13:35:47 General Forth info, sure, but what about a blog setting for people with specific projects? 13:38:44 --- quit: madgarden ("*poof*") 13:46:47 thin: I *finally* got the e-mail. :) 13:51:43 --- part: Speuler left #forth 13:53:51 heh 13:54:02 woops, afk for ab it.. 13:55:12