Logging Started : Sat Jan 31 13:43:54 EST 1998
graydon:	are you on the mailing list or just here out of curiosity?
vertigo_:	hi!
vertigo_:	i'm lurking on the mailinglist just out of general interest..
vertigo_:	and just saw your mail coming in about this meeting so i thought it might be interesting :)
graydon:	cool. This meeting is logged, feel free to ask questions..
vertigo_:	okidoki :)
vertigo_:	are you expecting the majority of the list to pop-up here?
graydon:	johnny (ANOQ of the sun) will not be attending. I think loren and jonas will be here, hopefully mentalboy, paul, jordan, and some others will drop by, but I really don't know.
vertigo_:	ah.. ohwell we'll see then :)
graydon:	how long have you been on the list?
vertigo_:	i think an irc meeting is a good plan to get things a bit constructive again after all the fuss lately.
vertigo_:	don't know, month or 2, 3 i guess
vertigo_:	not that long
graydon:	I agree totally. People are feeling a little discouraged with that fighting, and we need to keep the enthusiasm.
graydon:	I got access to debian's machines this morning, so i am thinking I can put the web site back up.
vertigo_:	ah, thats a step in the right direction..
graydon:	are you interested in developing for the project? -- i.e. do you speak c/c++?
vertigo_:	i'm afraid not, i'm still learning c and the ins and outs of linux (only installed it for about 5 months now - allready grew to hate X11 in that period ;) but it can't hurt to monitor some projects like this and the ggi project..
graydon:	of course, you're welcome to do that ;)
vertigo_:	hehe :)
vertigo_:	it can't hurt (i hope ;)
graydon:	not at all. the more people know about our efforts, the more scrutiny and input we get, the more we are assured of success.
	jones (~jones@195.192.126.200) has joined channel #Berlin
jones:	hello all
vertigo_:	there's something i was wondering about btw, with a project like gnome going on at the moment that is bound to produce lots of programs, isn't there a possibility for berlin to make use of that in one or another way?
jones:	vertigo_: in what way? Source compatibility? A rewrite IMHO "must"/will still be required
vertigo_:	jones: at least some easy conversion or something, i don't know the technical implications of such a thing but it seems silly to me to ignore a large base of applications thats out there waiting, thats something that could possibly make or break the acceptance of a new system.
jones:	Sure it would be nice, but honestly, I don't wan't to design a system that in _any_ way comprimises _anything_ for compatibility! We've got the chance to develop something really new here, let's not tie us self to the past (although we'll have to look at the past ;)
	jones sends a BIG "WAKEUP GRAYDON!"-card to Graydon :)
graydon:	heh. awake now..
graydon:	sorry.
jones:	hi
vertigo_:	jones: grin, yeah i understand the motivation :)
jones:	How long to the "official" meeting time?
graydon:	to comment: most programs are easily ported.. or at least, it's possible, especially with free soruce. Have you seen how fast poeple move KDE apps over to GNOME? I don't think it's going to be too hard to port to berlin.
graydon:	jones: I don't know ;) I don't think we settled on a time because we're all in different countries!
jones:	right, sure the GUI might/will require some work, but hopefully most of the application will be "independant".
jones:	graydon: Didn't you write about the time GMT?
jones:	GMT+7 or something?
graydon:	Yeah, but then someone else was posting something about time in EST, which is actually where I am... I don't know. I just sent everyone an email, so we can hope they're awake and at a terminal..
	jones thinks it's a shame that Matt isn't here in the time of Jim's quit!
jones:	:)
	graydon agrees. The debian guys were a little squirly about who is running the project, but I told them it'll be OK, and that we both have a leader and lots of interested developers. They seemed to be OK with that.
jones:	So I think the topic for tonight will be organization (dns, webpage, etc) and _general_ Berlin design (relating core,warzaw and Moscow)
graydon:	Yup.
graydon:	jones: what time is it where you are anyway?
jones:	yeah, I gotta say I got quite disaapointed with Bruce's mail. Seems like he was doing this as a Friends-favor to Jim (butting out when he did :( )
jones:	time is 07:12 pm
graydon:	taylorcc says people are meeting on efnet, #berlin. I'm going to go check that out.
	Reconnected as graydon at Sat Jan 31 14:10:15 EST 1998
	vtg (predator@alumina.xs4all.nl) has joined channel #Berlin
vtg:	hrm.
	vtg has left channel #berlin
graydon:	is this a channel for berlin-project developers?
graydon:	I'll take that as a no.
	Reconnected as graydon at Sat Jan 31 14:11:56 EST 1998
	jones (~jones@195.192.126.200) has joined channel #Berlin
	vertigo_ (predator@alumina.xs4all.nl) has joined channel #Berlin
vertigo_:	ah..
jones:	heh, Graydon made it back b4 :)
vertigo_:	grin
jones:	so, was people on efnet?
vertigo_:	i'm used to the evil worlds of ircnet and efnet :)
graydon:	nope, just a bunch of MIxeD-cAPitAlizAtioN-DudeZ
jones:	We R S0 3leI7 :)
vertigo_:	ph34r BuHrL1n
	vertigo_ *shivers*
graydon:	jones: I'd like to meet bruce. He does a *lot* of linux advocacy.. I
jones:	I was on IRC yesterday, seems like a new script was out, because I was notified about HACK: all the time :( - the kids are playing :)
	graydon shivers double..
jones:	Enough, anyone know anyone else who's gonna join? Or shall we start discussing?
graydon:	The world's first self-aware AI will be developed as a mIRC script. and it will be totally incomprehensible.
jones:	vertigo - where in the holy land of Netherlands?
vertigo_:	jones: Leiden..
jones:	speaking of, was talking to someone who was talking about IRC as _mIRC_ - kinda freaked when I asked him if one was using another client :)
	graydon has set the topic: "berlin development: logistics (sigh)"
	graydon has set the topic: berlin development: logistics (sigh)
jones:	Leiden - the "small" city between A-dam and R-dam? Or am I misplacing it?
	ShadowX (steve@maxt3m19.ipoline.com) has joined channel #Berlin
vertigo_:	jones: grin, "there are no people today in mIRC" :)
ShadowX:	hello
	siliconma (don@brndas1b-p14.mts.net) has joined channel #Berlin
jones:	:)
graydon:	Whoo! peepls
siliconma:	Hiya, how's things
vertigo_:	jones: yeah thats correct, it's maybe 15km's next to The Hague :)
jones:	hello, stand up and identify yourself :)
	ixx (~ixx@accs04-9.ttu.edu) has joined channel #Berlin
ShadowX:	i'm just lurker on berlin list
ShadowX:	steve@ggi-project.org
siliconma:	Hiya, you know me from the berlin list as that guy with the stupit tasket idea
jones:	Went visiting a friend from the LIP project last summer in A-dam, damn I _LOVE_ that city!
ixx:	hello
	ixx == taylorcc
siliconma:	Graydon: I visited your pages, are you attending the U of Toronto?
graydon:	ok, the situation is this: SPI has given me an account on the debian master site, and will shortly be giving me access to a CVS tree and a web server.. so I'd like to get (in this log) a listing of all the people who have relevant things they've written / drawn / coded / archived, so I can copy them to the new site.
jones:	siliconmage: Love the idea, perhaps you could refresh my mind, on _exactly_ what it is! I got the impression it a higher/lower level of interaction allowing information to be displaced on screen/braile/sound etc, other people think it's something completely diff.
graydon:	soliconma: yup, math and comp-sci. This is my first year, although I'm older than my classmates (20)
vertigo_:	jones: yep, it's a nice place to go out, but i wouldn't want to live there. lots of criminality and stuff :/ leiden is as peaceful as it gets.. and only 20 minutes by train from a'dam ;)
siliconma:	graydon: cool, happen to know a master's student names Neil Enns there?
jones:	vertigo_: nice, althought I did find adam peacefull :)
graydon:	silicon: U of T has 50,000 students. unless he's in first year math, I wouldn't have met him.
jones:	graydon: nothing yet, but def. planning on :)
siliconma:	jones: I'm working on a web page describing it, I'll email the list when I'm done writing
siliconma:	graydon: Naw, he's almost done his master's in CS working on HCI
jones:	Ok, people are joining :) Let's wait another 5 min. before starting
graydon:	jones: sounds fair..
	Derrida (~kclark@dal284.cmpu.net) has joined channel #Berlin
graydon:	just to let people know: I compiled the base skeletons for the warsaw server library last night (not the implementation, just the server runtime library) and it weighs in about 400k.. the omniORB library is about 1 meg on top of that.
ixx:	ick!
jones:	graydon: I was thinking about writing a corba comp. that extracts events from X and distributes them, know how I can have the server object do something like that? Seems like boa->impl_is_ready() :(
jones:	graydon: did any -O ? or -g?
graydon:	jones: no optimization, no debugging symbols, no stripping. Just what comes out of G++ -- of course, that's got a lot of concurrency handling in it. it's completely reentrant.
ixx:	ok
siliconma:	Anyone know what Jim's position re leaving ended up?
jones:	btw - seems like gcc 2.8 still doesn't handle namespaces :(
graydon:	silicon: haven't heard back from him. AFAIK he left the list, but he'll be back someday, I bet.
Derrida:	jones: what about egcs?
jones:	siliconmage: I gotta say I'm disappointed, first it seemed like he had leader capabilities, but his "running away" leaves me with the impression that he's def. not!
graydon:	not namespaces, but it does do proper, safe C++ exception handling across shared libraries, which is quite an improvement.
Derrida:	jim left in amazingly irresponsible way...why wasn't he sharing with everyone else his discussions with SPI?
jones:	Derrida: Tried it for a while, but found the c++ part bad (coulnd't compile LIP, which gcc handles fine!) - poor template handling!
jones:	graydon: that problem I've had! Thank god they fixed it!
jones:	Derrida: some mails to the list, don't think people were paying much attention
vertigo_:	i don't want to sound harsh, and maybe it's a loss of somebody who really knows his stuff, but i can't help to think that jim's attitude sometimes has also scared away a lot of potential developers and/or interested people.
graydon:	jones: NOTE: egcs 1.0 is not compatible with egcs 1.01 and g++ 2.8.0 -- egcs 1.0 handles exceptions differently.
jones:	vertigo_: haven't seen anyone leave because of it, but his mail regarding Graydon, really got me off!
jones:	not off the list ofcourse ;)
vertigo_:	jones :)
jones:	graydon: did just remove egcs 1.0 after installing gcc 2.8 last night
Derrida:	i was surprised by how much Jim seemed to dislike the hurly burly of free software development...I think he wanted to build a cathedral, not a bazaar
jones:	Derrida: I really like that model :) - still we need coordinators though - believe me
graydon:	jones: the way I see it, jim's not my boss, berlin doesn't pay my bills. It's a hobby, and an incredibly fun one, so I'm going to write this software out of personal intrest, not whether I have his approval or support. I'm sorry he left though, because he knows color theory backwards, and I don't know it at all.
jones:	the LIP has / have no coordinators and often it's hard to know what people are working on! Only when people knows each other, can it work - and still not to it's full
jones:	graydon: Exactly my words :) - he was skilled, but perhaps not "built" for bazaar :(
Derrida:	jones: we've got coordinators aplenty...imho
jones:	Also, I find it quite frustrating Matt isn't here :( - Seems like whenever the core gets mentioned the response is: Matt will do it. Since his not been heard from for a long time :(
Derrida:	graydon is right: code for the fun of it, try to keep in touch with the others who are coding, and keep digging!
siliconma:	everyone: I'm wondering what I can do here... so far all I've been able to do is contribute a little to the email discussion. What needs doing??
	graydon has set the topic: "berlin development: How To Contribute!"
	graydon has set the topic: berlin development: How To Contribute!
siliconma:	:)
jones:	Derrida: But where is the listing? Matt = core, Graydon = warzaw, Moscow = Not ready yet?
Derrida:	i volunteered during jim's 'meltdown' to maintain the website, but he never answered me;
Derrida:	i'd be willing to do that still, so that a programmer doesn't have to
Derrida:	jones: sounds good for right now
jones:	siliconmage: It would be nice to write API! But please, finish up your tasket idea first! I find this _very_ important!
jones:	graydon: You're taking the "job" as coordinator for Warzaw?
graydon:	if you're going to contribute code, you need your system at home configured with GGI 0.0.9, G++ 2.8.0, libg++ 2.8.0, and omniORB 2.4.0... absolutely everyone needs this.
siliconma:	ok, are we looking for a vauge "this is what a tasket is" or would you like me to take a stab at an api already? (I can answer that... both! )
jones:	graydon: is omniORB on with the "general" naming scheme? Object_impl : public Object_skel ?
jones:	siliconmage: please start by describing the IDEA - what exactly is a tasket!
graydon:	jones: there's no standard for that *precise* part of server-side implementation, with omni orb it's Object_impl : public _sk_Object
jones:	siliconmage: You can wait with the API, as far as I know Graydon, has started doing Taskets api with Warzaw
graydon:	close enough that is only takes a minute to switch ORBs.
jones:	graydon: I know (not until corba 3.0?), but most code I've seen used Object_skel - not really _that_ importatn
graydon:	wrt taskets: I can only offer this idea.. Imagine you have a widget, something which is just like a widget, except there's more than one obvious way to do what it does, which makes no difference to the application. It's just a *little* more abstract than a widget, but enough that you add this extra layer of abstraction.
graydon:	for instance, mutually exclusive selection from a number of options can be done both with a select list, a combo-box, or a radio group. Which do you use? you use the abstract form, and let some level of user preferences decide for you.
siliconma:	graydon: you understand correctly.
jones:	graydon: I makes it kinda hard to handle screen-real estate + developers won't have full control (ofcourse users may benefit:)
graydon:	Likewise, you might decide to do it through a speech interface, in which case there aren't even accompanying visuals at all.
jones:	it hurts in my head thinking about how to do it :)
ShadowX:	lastlog 50
jones:	graydon: perhaps every widget _shouldn't_ be derived from tasket, but rather have tasket-groups?
graydon:	jones: nope, you're forgetting that widgets are beneath taskets in the class hierarchy, and all taskets are abstract, so you don't actually instantiate a tasket -- you always have a widget of some sort, it's just that the only *type* information you know if that it's of some tasket type. it will still obey screen layout methods.
	Derrida is away: (Auto-Away after 10 mins) [BX-MsgLog On]
jones:	And perhaps have these taskets-groups at a higher level?
graydon:	a tasket is really just a logical grouping of related widgets.
	graydon has set the topic: "berlin development: oh god, we're discussing taskets again :)"
	graydon has set the topic: berlin development: oh god, we're discussing taskets again :)
siliconma:	As I see it a tasket is a way of runtime binding an app to a widget with the widget selected for a particular task being under the control of the user
jones:	graydon: I was just thinking about more groups os taskets (input tasket, mutex tasket, text tasket, output, etc) instead of one ( widget : public tasket )?
graydon:	hmm, jones: of course, what the hell was I thinking?
graydon:	oh..
jones:	Can we (please) save the tasket discussion for either later this evening or a later warzaw meeting? - still got plenty of other topics!
siliconma:	k
graydon:	ok, interface tasket is the point in the class hierarchy where the reactor, observer, etc. classes all congregate.. ok, I'm less confused by my own reasoning than I was.
graydon:	jones: yeah, lets do that topic some other day.
jones:	ok, I've written a small "agenda" of topics, mayor topics are: Organizion, Berlin overall design.
	graydon has set the topic: "berlin development: Organization (sounds official huh?)"
	graydon has set the topic: berlin development: Organization (sounds official huh?)
jones:	graydont: topic = organization - subtopic = Domain name - how do we hanle?
graydon:	Who owns the domain name?
	graydon has set the topic: "berlin development: Organization -subtopic: domain name"
	graydon has set the topic: berlin development: Organization -subtopic: domain name
graydon:	anyone, anyone?
graydon:	ok, presumably whoever owns it is not here.
graydon:	I'm guessing it's jordy. Jordan has an ISP, so I think he's hosting it.
graydon:	We'll ask him to point it to the SPI site as soon as I've moved a respectable amount of stuff there.
jones:	ok, topic set. How do we handle it? I mean payment, machine etc, now that Jim is gone :)
graydon:	I'll mail the guy who runs the SPI site and ask him to set aside a redirect.
jones:	SPI == Debian ? - Sorry, haven't done my homework here :(
graydon:	Yes. SPI == "software in the public intrest" == Debian's owning nonprofit organization, which primarily focuses on making Debian.
jones:	Ok, Jordan handles it. What about payment? Probably isn't free?
graydon:	No, domain names cost $100 to make (paid for already) and $50 annually to maintain. I can spring for that, it's not too bad.
graydon:	Between jordan and myself I think we have this topic nailed down.
jones:	graydon: GREAT!
	graydon has set the topic: "berlin development: Organization -subtopic: who has stuff for the web site?"
	graydon has set the topic: berlin development: Organization -subtopic: who has stuff for the web site?
jones:	ok, next topic - website!
jones:	We need whitepapers! - and a restructuring!?
siliconma:	ok, stop.
siliconma:	whois does not list that jordy has much to do with berlin-consorium.org
siliconma:	(consorium)
siliconma:	Administrative Contact:
siliconma:	Messier, Matt MM162 marius@PRILNARI.COM
siliconma:	401-726-4828
siliconma:	Technical Contact, Zone Contact:
siliconma:	Kail, Jamie JK391 pyro@PYROTECHNICS.COM
siliconma:	(847)991-9345
siliconma:	Billing Contact:
siliconma:	Fetters, Jim JF4320 jfetters@PRILNARI.COM
siliconma:	602-246-8936
siliconma:	does that change anything?
graydon:	Jordy has indicated to me that he has been in contact with Matt (I have seen Matt post messages, so I *KNOW* the man exists ;) and can get the domain name pointed to the right place. He may be yanking my chain, but I trust him on this issue.
siliconma:	k
graydon:	I thought he was also actually serving it off his home site, but maybe not. lemme play with nslookup for a second..
jones:	graydon: he has a mirror, yes
graydon:	name service is coming from pyrotechnics, yeah. I don't know who Jamie Kail is... anyone know?
siliconma:	nope
jones:	graydon: hasn't the name pyro showed on the list from tie to time?
graydon:	jones: no idea. Ok, well, when I'm done with this meeting I'm going to start uploading what I have to the SPI site, as well as (hopefully) getting jordy to send me his mirror of the berlin site. When that is going down, I'll talk to him about the domain name. satisfactory?
	Signoff: ShadowX (Leaving)
jones:	graydon: ok
jones:	ok, website - how can we enhance it?
graydon:	jones: I don't mean to be horribly critical of the old site, but I'd like to ... compact it a little. Make it more technically dense and easy to navigate.
siliconma:	Well, I think the question is how's attention do we want to attract?
siliconma:	(how's->who's)
graydon:	I think linux folks are hard to impress, and you need serious content available up front. Like, a link on the front page to browse the source.
siliconma:	If we want more technical people then we should have a site which appeals to them
jones:	graydon: exactly, more techie stuff (for now!) and clean-up!
siliconma:	I.e. dense technical info on the project, it's easy to find how to get on the mailing lists, etc
graydon:	Yes, and how to grab the CVS tree (CVS is, like, god. if anyone here hasn't used it -- learn how!)
jones:	graydon :) - maybe not that "hardcore", but close. I know the ggi has Browse Source :)
siliconma:	I use it at work, wish they'd taught us about it in class!
graydon:	jones: yeah, I mean, I *really* like the graphic design emphasis on the site, but it also makes a lot of sorta unsubstantiated claims with mockups..
jones:	cvs rules! How about authentication, etc? Who will admin?
graydon:	It would be more convincing if, side by side with the FAQs and mission statements and all, there were architectural drawings and explanations.
jones:	graydon: exactly, the graphics are nice, but the site is quite difficult to navigate, and once you find it, it's like a screenshot explaining nothing about the "internal workings" - read: WE NEED PAPERS!
ixx:	oh hey.. i can offer space for berlin "stuff" on my server
graydon:	jones: martin schulze, who is a debian maintainer-maintainer, says we have access to CVS. I'm not sure how that will work, but I'm going to find out very soon cause we live by CVS at work and it helps a whole lot.
siliconma:	Question: is there only the 6 of us who are really interested in this... or did some people not show?
ixx:	right now i am coding in ggi... well actuallly trying to port something to ggi.....
jones:	I can make a european/danish cvs mirror
jones:	siliconmage: we've had problems with coordinating time :( and servers :(
graydon:	there are others. I think it would be awesome if everyone who is interested and/or on the list could contribute a little summary of who they are, what their intrests are, what they're working on, etc.
ixx:	i have been on the berlin list forever.... don't know when i'll have time to do anything other than toss ideas though
ixx:	but i can offer ftp, cvs, web etc
ixx:	and mail accounts
graydon:	ixx: full time?
jones:	Pers. info - look at lin01.vejlehs.dk/~jones - still need a picture and some more info :)
ixx:	graydon, fulltime? what do u mean?
jones:	ixx: Ideas are very important - otherwise we'ed just be coying the present :)
graydon:	ixx: I mean.. you've got a machine with fulltime access to the net?
jones:	I've got machine with fulltime access (lin01). Not at home thought :)
ixx:	yes
ixx:	its at my unv. in a proffesors office
graydon:	cool. That's 2 mirrors, as far as I'm concerned. Excelent.
ixx:	connected via ethernet on campus.....
siliconma:	speed!!
ixx:	we have 9t1s going out from the school
ixx:	if i do code for berlin, i would like to work in the core if possible
jones:	ixx: Matt's doing that :) - *JOKE*
ixx:	:)
graydon:	heh, ok folks, we gotta get this moving along. Let me put the web site situation this way: SPI is really really cautious about people getting access to their site, but if you guys have external sites, you can put things there too. When I know more about how CVS is going, I'll port to the list, until then, assume I"m working on getting the site and DNS switched over. OK?
graydon:	port=post.
graydon:	ixx: what's the name / address of the site you have access to?
graydon:	hello?
	Reconnected as graydon at Sat Jan 31 15:23:33 EST 1998
graydon:	zoiks! everyone still here?
jones:	yuup
graydon:	ok. Lets move along. next topic?
jones:	Handling SPI?
	graydon has set the topic: "berlin development: Organization -subtopic: SPI"
	graydon has set the topic: berlin development: Organization -subtopic: SPI
jones:	Should we give berlin a couple of months to mature, before proceeding?
graydon:	jonesL what do you mean by "proceeding" and "mature"? .. progress is occuring.
siliconma:	I'm curious as to whay SPI can do for us... i.e. what is the benefit of the relationship?
jones:	before proceeding with "publishing" berlin, perhaps we should have some more stable ground (papers and source)?
graydon:	silicon: they have tremendous publicity and presence in free software. They are the second largest linux distribution. The president of RH talks to bruce a lot -- they're very good friends to have.
jones:	siliconmage: me too, one benefit I've heard about is that Bruce knows RH pretty good
ixx:	$$$
ixx:	and image
jones:	not only $$$ - we need people like RH and debian if we _really_ want success!
ixx:	and rh yes
ixx:	yes
siliconma:	So i assume we eventually want to become part of the standard linux distibutions
ixx:	of course :)
graydon:	RH is basically freewheeling in this business -- they'll take anything they think is promising and promote it, hoping to put their brand name on it. To be honest, this is fine with me, if they can promote berlin (when it's somewhat more .. done .. than it is now) I'm all for it.
jones:	siliconmage: hopefully, I'll be happy when people can choose between X and Berlin and debian and RH!
	derry (~derry@large.reno.nv.us) has joined channel #Berlin
jones:	ok, so we wait on contacting "people" before we have some more. Perhaps a server displaying simple windows?
graydon:	The thing that makes SPI so useful is that thay have a less commercial feel to them. they're nonprofit, they make sure everything which has non-free copyrights is separate in their work.. this gives them a lot of credibility with linux coders who are afraid of commercialization.
graydon:	jones: yeah, you were talking about some kind of demo.
jones:	yes, A small server showing x (where x is a number :) windows, perhaps showing unicode and registar, would be a good starting base
Derrida:	graydon: do all the ORBs you're looking at come 'unencumbered'?
graydon:	derrida: unencumbered by restrictive copyrights? Yes. Mico and OmniORB are the contenders (plus TAO if I can ever get it working) and all of them are 100% GPL.
	graydon has set the topic: "berlin development: Publicity -- a demo"
	graydon has set the topic: berlin development: Publicity -- a demo
siliconma:	I've got to go guys, I'll see you on the list.
graydon:	silicon: bye. have a nice day
	Signoff: siliconma (Leaving)
jones:	graydon: do you know what kind of services they plan on? There a lot of services (transaction, events, persistence, etc)=
graydon:	jones: name service, a little lifecycle service. not much else.
jones:	graydon: do none of them plan on implementing more?
graydon:	jones: I think a good demo to aim for is opening windows which have the freetype equivalent of an xterm.
graydon:	I don't think many more services are standard.. TAO has *lots* of services, with real-time extensions too, but I can't get it to work.
jones:	ok, so we agree that the first "demo" should contain an amount of windows, atleast one shows an "xterm"
graydon:	jones: yeah, that would be an amazing demo. It's just .. it's a little ways off. Lets make a list of things which do *not* need to be in such a demo.
graydon:	* no registrar needs to be in the demo
graydon:	(like, nothing persistent)
jones:	unicode should! Show iso8859-1 and japanese output!
graydon:	* no high-performance optimizations
graydon:	jones: YEAH! I'd love to get japanese truetype fonts. A tricky operation -- that vector library I was pointed towards might have some, I never finished investigating that.
derry:	sorry to butt in the middle here, but what kind of demo are you guys discussing?
jones:	graydon: someone on the list mentioned a TT or Type1 japanese font with unicode ?
jones:	derry: a demo showing Berlin - to the public
derry:	jones: do you mean one that shows the idea/concept of berlin, or some real working code?
jones:	derry: _real_ working code
jones:	something that people can download and see that Berlin is more than tlk
jones:	s/tlk/talk
graydon:	jones: yeah, they mentionned a type 1. The thing is, I have managed to get freetype to render to a GGI, but I haven't got a type 1 rasterizer. jim was going to make one, but he never did.. if you can find a truetype font which is not 8859-1, we can use it.
Derrida:	exit
	Signoff: Derrida (Derrida has no reason)
jones:	heh, ixx - you wanted to work on the core? Font rasterization needs work :) - BAD!
ixx:	heh
graydon:	jones: well, only type1 fonts. I'm perfectly happy using freetype for the first while, letting someone (like, say, adobe) write the type1 section. I'm very lazy, you know..
ixx:	freetype does not do type1? hmmmmmm
ixx:	what are we lacking in freetype
graydon:	ixx: no, freetype does truetype fonts. They're an entirely different technology. Sadly. They're being merged in the "opentype" standard, but truetype is good enough for lots of applications.
graydon:	Especially screen display and non-commercial print, they're just fine. They scale, rotate, etc. They just use a simpler internal representation of curves.
jones:	graydon: did you put all the code for your freetype displayer online?
graydon:	jones: yeah. You actually only need to rewrite the graphics driver stub in the freetype demo directory (in the main freetype distribution). So I put the driver rewrite on line.
ixx:	i gota go install a new drive in gods... i'll come back later and see if you guys are still here
	Signoff: ixx (BitchX: the offical sponsor of the 2002 Olympic Winter Games.)
graydon:	I can only stay around for another.. maybe 10-15 minutes.
graydon:	sorry :(
graydon:	jones: do you know how to aquire a pty (my system-level C is a little vague here) in the manner that xterm does?
graydon:	Presumably we could adapt the rxvt code to fit our purposes... I'll go pick that up.
	MenTaLboY (~mental@207-172-233-5.s5.as1.wmn.erols.com) has joined channel #Berlin
MenTaLboY:	afternoon folks.
graydon:	I don't think the fonts will be the hard part -- I think getting windows on the screen at all, responsing to the mouse and keys, wil b the hard part.
graydon:	Perfect timing mentalboy!
	MenTaLboY grins
graydon:	just getting to your GWT code.
MenTaLboY:	I could sense it. :)
jones:	sorry, was away for a minutte.
graydon:	we're talking about the potential to make a demo (non-persistent) which opens up a screen and lets you open xterm-like creatures in little windows. So this is the time to discuss GWT. I haven't tried it yet. What does it do?
jones:	bterm - Allan from LIP wrote a LIP term, so I can ask him to help - I think he can be persuaded :)
MenTaLboY:	Not too much by itself.
MenTaLboY:	It just manages window stacks and invaid regions
MenTaLboY:	actually drawing is left to another library
MenTaLboY:	(i.e. LibGGI2D)
graydon:	jones: we must *not* get into the habit of calling everything we write b-xxx- (i.e. bterm, bpaint, bemacs)
MenTaLboY:	generally, the way to paint a window would be something along the lines of:
MenTaLboY:	graydon: indeed!
jones:	Speaking of, will all windowing in warzaw be corba objects? Or can we benefit from having a lower level lib and wrap it?
graydon:	mentalboy: that's fine, I have stuff which can draw -- rendering text is just copying scanlines out of freetype's draw buffer. it's easy.
MenTaLboY:	yep
MenTaLboY:	generally, you'd just get the window's region, then get it's 'invalid' region, and then it's visual
jones:	bterm - Hmm, I don't mind, But what about "Berlin Window to ley You command the system by keyboard"?
MenTaLboY:	(the invalid region would be relative to the window's region)
jones:	or "Madrid"? - ok, just open Madrid and type.. :)
MenTaLboY:	jones: no... let's not go down this prefix path again
graydon:	jones: don't know. I suspect we can use a light-weight C library for it, but I'll probably only know when I am up to my elbows in implementation.
MenTaLboY:	Anyone happen to know the german word for 'teletype'? If it's relatively short, it'll do nicely.
jones:	what about "fernschreiber"?
graydon:	jones: If we name everything after cities, nobody will know what the hell any programs do!
	MenTaLboY nods
MenTaLboY:	quite
MenTaLboY:	hrmn
graydon:	jones: shorten it: "scriber"
MenTaLboY:	I think I have a german dictionary around here someplace
jones:	My german isn't that "gut" but it's prob. something like that
jones:	vertigo_: What's it in dutch?
vertigo_:	don't know to be honest
MenTaLboY:	damn... I can't find it
derry:	how about something short like 'vt' (virtual terminal)
MenTaLboY:	yes
MenTaLboY:	that'll do
MenTaLboY:	well, maybe one more letter in ther
MenTaLboY:	wvt ?
jones:	anyway, let's not waste precious time on this (for now)
MenTaLboY:	or.. better ... 'vtw' ?
graydon:	in french: "two way telegraphy" --> "duplex"
derry:	yes, sadly we are working on a name before we have anything to name
MenTaLboY:	hrmn... actually, 'virtual terminal' has a specific meaning under linux that wouldn't be fulfilled by an xterm-type app
MenTaLboY:	maybe 'ttw'
MenTaLboY:	(teletype window)
jones:	people, were getting of track here :)
graydon:	jones: harumph, you're no fun.
graydon:	jones: sigh. ok.
jones:	:)
graydon:	Anyway That should be enough to keep us terirbly busy.
jones:	graydon: remember last meeting? We need to stop it at some time :)
	Schubert (~mschubert@usr-x2-hub2.ppp-34.bak.igalaxy.net) has joined channel #Berlin
Schubert:	hello
graydon:	hey
jones:	hello
MenTaLboY:	hello.
derry:	I must have missed a lot, what has been decided?
MenTaLboY:	not much, afaiK
jones:	everybody - I'm writing notes, these notes will be on my site when the meeting is over (lin01.vejlehs.dk/~jones/meeting.html)
graydon:	derry: I think the decision is that jones, mentalboy and I are going to work on a demo integrating corba, GWT, and freetype, to give a rudimentary terminal emulator.
jones:	We've discussed organization - dns, webpage, etc
MenTaLboY:	ahhh.. here we go
graydon:	we were also going to discuss moscow. Does anyone have *any* clues on that issue yet, or is it premature?
derry:	ok. I would really like to get involved in this. What can I do?
MenTaLboY:	here's my post to the GGI mailing list about GWT: http://www.ggi-project.org/mailinglist/jan98/1522.html
jones:	graydon: I've got it in my list of topics
MenTaLboY:	Moscow would be stuff like clipboard and drag-and-drop constructs
MenTaLboY:	and various other services that would be useful for shell apps
jones:	MenTaLboY: I don't agree (or maybe I do?) but I see Moscow as the _desktop_ like windows explorer
graydon:	derry: first of all get yourself set up with the compiler(g++2.8), library,(libg++2.8) GGI 0.0.9, and omniORB 2.4.. It takes a long time (sadly)
MenTaLboY:	no, PLEASE let's not confuse the operating environment with the shell
MenTaLboY:	as windows has done... for one thing, it's too inflexible
Schubert:	Well I just dropped by, I've got to go somewhere, when I get back, I'll start working on a cleaner version of the mailing list
jones:	Moscow is the operating environment then. Atleast that's my view of it
graydon:	derry: at the same time, if you want to dissect the rxvt sources, learn how they work, that would be very keen.
derry:	graydon: that I can do, how about the current source code?
	MenTaLboY nods
MenTaLboY:	jones: okay, I'll agree with that much
Schubert:	hmmm
graydon:	schubert: I looked at it today. Very useful! I'll put a link (and/or mirror) on the new belrin page.
MenTaLboY:	jones: I just think users should be able to use different shells (and thus entirely different desktops) under Berlin, if they so choose
jones:	We need an operating environment! (Moscow to me).
MenTaLboY:	Moscow would provide the sort of services such shell applications would need, thus the operating environment
jones:	Different desktops are no problem! Look at windows - more than one desktop exists.
Schubert:	be back later...
graydon:	derry: you can read over the corba tutorial I put up.. it's pretty much the same story under omniORB as it is under mico. http://www.pobox.com/~graydon/corba_intro.html
MenTaLboY:	yes, but I've done enough Win32 programming to know that a lot of OS functionality is tied up in Windows Explorer, which is just downright wrong
jones:	I was thinking about Moscow as (again) an idl definition. Providing services, then different people can implement it to their liking?
MenTaLboY:	bingo
MenTaLboY:	that's good
	smkl (sami@ppp7.dial-in.vtoy.fi) has joined channel #Berlin
MenTaLboY:	that also kinda fits in with the clipboard ideas I suggested on the ML recently
graydon:	derry: I also have sources on my page for the GGI modification to freetype, if you feel like playing with freetype. It's not too difficult though, I doubt it'll occupy much of our time.
MenTaLboY:	hey
derry:	I will look into rxvt. I have looked at CORBA and your stuff. Is there a current snapshot of berlin available someplace?
jones:	moscow: By being a definition my way of executing programs can be very different from my neighbours (assuming he is running Berlin :)
MenTaLboY:	the berlin core?
MenTaLboY:	I think jim had the archive
MenTaLboY:	jones: that's the general idea
jones:	the core - get it via cvs? I got mine that way
MenTaLboY:	jones: that will allow the same sort of flexibilty people have now with choosing various login shells
MenTaLboY:	jones: definitely good
jones:	MenTaLboY: then we agree :)
MenTaLboY:	jones: yep
graydon:	derry: not really. that will become more clear as we move into the space SPI gave us. Berlin, as a "thing", is still really in pieces. there's the core library, if you want to look at the registrar and unicode stuff, and I put the GWT code mentalboy did on my warsaw page..
	MenTaLboY grimaces
MenTaLboY:	ungh
MenTaLboY:	that code is barely ready for primetime
derry:	So I can get it off your page, graydon?
jones:	ok, graydon can you stay for how long?
MenTaLboY:	say, has anyone read http://www.ggi-project.org/mailinglist/jan98/1522.html yet?
graydon:	derry: you can get mine and mentalboy's. I don't know if the CVS server with the core is still up.. I forget it's address. JOnes, you know?
graydon:	jones: well actually pressure's off, I can stay a bit longer, cause my girlfriend has decided to start making cookies.
jones:	I think it would be good with a coordinator - and people sticking to the subject
	MenTaLboY is now known as YMenTaLbo
YMenTaLbo:	gahh
	YMenTaLbo is now known as MenTaLboY
	jones would love a homebaked cookie ;)
MenTaLboY:	too bad you can't DCC baked goods...
derry:	I have a web and ftp server available if that would be helpful
graydon:	mentalboy: even if your code isn't ready for "prime time" all it needs is the encapsulation of an idea, and others can work on it. The problem is when you publish totally *aimless* code, it's not useful. But if it has a clear intent, it's worth publishing.
derry:	yeah, homemade cookies sound good
MenTaLboY:	graydon: I guess... at least the intent is covered
jones:	ok, the core - I'm tar /gzipping mine now, and putting it on ftp://lin01.vejlehs.dk/pub/Berlin/core.tar.gz
MenTaLboY:	heaven knows I've got enough /* FIXMEs */ in there specifying desired behaviour
graydon:	jones: excellent. I'll grab a copy and try to incorporate it into my uploads to SPI..
MenTaLboY:	ungh
MenTaLboY:	just realized that the private data functions aren't threadsafed
MenTaLboY:	brb
derry:	jones: thanks I'll grab it
graydon:	mentalboy: That's OK. You need only put threadsafing locks on something if you suspect it will be accessed simultaneously from multiple threads. If we encapsulate window management in a single thread (which I think is perfectly reasonable) then we don't need to modify your design.
MenTaLboY:	graydon: well, just since I've threadsafed everything else... :)
MenTaLboY:	graydon: anyhow, LibGWT is more general than just for Berlin -- ultimately it's going to become part of the GGI distribution
graydon:	BTW: no matter what ORB we use, I encourage everyone to use the omniThread library which shippes with omniORB. It's a very small wrapper that lets you do C++ threads the way you do them in java: i.e. subclass from thread, and you get an object with thread-specific data, which can spawn into its own personal method. This is *WAY* better than the pthreads interface.
	MenTaLboY nods
MenTaLboY:	oh... one more issue
jones:	ok, the berlin core is ready for download. ftp://lin01.vejlehs.dk/pub/Berlin/core.tar.gz
MenTaLboY:	how much thought have we given so far to Berlin on PDAs and such?
graydon:	mentalboy: ok.. threadsafe away then!
	Loren (lso8219@grace.isc.rit.edu) has joined channel #Berlin
Loren:	brb
	Signoff: Loren (Leaving)
MenTaLboY:	heh.
vertigo_:	now that was fast :)
	Loren (~lso8219@grace.isc.rit.edu) has joined channel #Berlin
MenTaLboY:	so was that
jones:	ehhh, I didn't even see him comming - he's back :)
jones:	hi loren
MenTaLboY:	afternoon Loren
Loren:	hold on.. my terminal is messed!!!
	Signoff: Loren (Leaving)
graydon:	mentalboy: I'm very interested in the PDA angle, but someone would need to publically declare support for free software, and ELKS would need to come a long way from where it is.
jones:	PDA = handheld pc's or am I off? ELKS = ?
	Loren (~lso8219@grace.isc.rit.edu) has joined channel #Berlin
Loren:	whoa... I can see.... :)
graydon:	jones: yeah, PDA == personal digital assistant, ELKS = version of linux for them. Don't cringe, it's just another fun idea.
graydon:	"embeddable linux kernel subset" .. it just barely works.
vertigo_:	taking linux everywhere you go sounds like the ultimate geek-experience ;)
	Schubert is away: (Auto-Away after 10 mins) [BX-MsgLog On]
Loren:	graydon: seen Brother's new laptop?? it's sortof neat... no HD though... runs GEOS...
MenTaLboY:	graydon: there's been some discussion on the GGI mailing list about installing Linux-GGI on larger PDAs ... some guy works for a company that makes them...
	graydon has set the topic: "berlin development: we are *SO* off topic!"
	graydon has set the topic: berlin development: we are *SO* off topic!
MenTaLboY:	I remember GEOS...
Loren:	anyone know how to make a winNT terminal behave...
Loren:	???
Loren:	mine's not...
MenTaLboY:	install Linux and use xterm.
	MenTaLboY grins
Schubert:	back... (for now)
Loren:	Mentalboy: not my machine... I'm at work...
MenTaLboY:	good luck then...
derry:	loren: you mean telnet?
Loren:	derry: yep...
vertigo_:	loren: try crt 2.0, it's a pretty good win32 telnet/terminal proggy
jones:	-STOP- can we please set topic, discuss topic, topic = topic->next?
derry:	loren: I think its just screwed, no way to make it behave
MenTaLboY:	I really think the Win32 distinction between 'console' and 'window' processes is really braindead ... everyone can benefit from stdout and especilaly stderr
Loren:	It's not auto-scrolling like it should... the bottom line just keeps getting overwritten and the screen doesn't scroll...
MenTaLboY:	Loren: no real way around that if it happens, afaik
MenTaLboY:	oh yeah... CRT is good stuff
vertigo_:	loren: not some line-feed setting or something?
MenTaLboY:	most everyone I know who uses Windows uses it
Loren:	I think it's fixable with tty?? no?
Loren:	hold.. I'll try to fix it..
Loren:	hello?
MenTaLboY:	WinNT telnet is just vt101 or vt102 emulation
MenTaLboY:	albeit broken
Loren:	Ahh... success... :)
MenTaLboY:	try each
graydon:	jones: what's the next topic?
Loren:	okay... how late am i???
jones:	Loren: very late :)
jones:	I was thinking about biG - where does it fit in?
	graydon has set the topic: "berlin development: big and the core"
	graydon has set the topic: berlin development: big and the core
MenTaLboY:	BiG makes a nice minimalist toolkit for working with 24bpp+alpha images, which is supposed to be Berlin's 'native' internal representation
MenTaLboY:	however, some issues have been raised about that...
MenTaLboY:	a lot of graphical-type people like to have 48bpp color
Loren:	what is big's job exactly?? to abstract all video modes to 24+8bpp???
MenTaLboY:	Loren: not ... quite
Loren:	Mentalboy.. then: do tell
MenTaLboY:	Loren: it's just that Berlin as previously envisioned by Jim would always use 24+8bpp for images, and reduction to the display's mode would be done transparently by Berlin
graydon:	mentalboy: I agree. I was thinking this over this morning at breakfast, and realized we're going to need a way to tell if a visual (a GGI visual) is on the current server or remote..
Loren:	Mentalboy: Berlin will?? what part of berlin will?
MenTaLboY:	when the images actually got displayed by the Berlin server is the only time that any conversion would happen, and then it would be invisible to the berlin apps
MenTaLboY:	Loren: the server, most likely
Loren:	graydon: why?? what does it matter on which machine we are???
MenTaLboY:	Calm down, Loren.
Loren:	sorry
graydon:	metalboy: I think it's a good idea to do some things in 24bpp.. My argument with jim was that if you have non-color-altering operations to do, like just *moving* pixels already in a display buffer, there's no point taking them to 24 bit and pack again.
	Loren will shut up and listen now...
	jones suffers from process overflow - doing IRC and cooking and tv is showing Deep Space Nine :)
MenTaLboY:	graydon: I'm starting to think that way too now, especially since restricting it to 24 bpp would piss off some people who like to be able to use 48bpp color
graydon:	the question, afaik, stems from a lack of understanding on my part as to how GGI actually tells which visuals use which functions..
graydon:	like, if I have a memory-target visual, it's on system memory, so I will necessarily have to copy it to get it to a screen, right?
jones:	a really liked the idea about RGB+A because it gives a consistent interface. But there are a lot of problems, with respect to HW accell, and higher bpp
Loren:	Mentalboy: what I mean is the Berlin server has lots of pieces... which piece will do the abstraction to 24+8bpp???
MenTaLboY:	Loren: probably whichever piece actually blitted the 24+8bpp to the display mem.
graydon:	but if I have a visual in the display buffer at the moment, it's a different call to move it.
	MenTaLboY nods
graydon:	mentalboy: yeah, maybe if we have a sorta staggered architecture, where the applications *think* everything is a 24bpp image,
MenTaLboY:	when a visual is created, the 'display target' is specified in the ggiOpen() call
graydon:	and then, if the library is actually displaying on a 24bpp card, it'll use hw accel, otherwise it'll fake it in-memory and then downsample the color
MenTaLboY:	from there, the display target code determines which functions to load into that visual
derry:	what advantages does 48bpp have?
MenTaLboY:	some people can actually see differences in 24bpp color
	Schubert is back from the dead. Gone 0 hrs 13 min 38 secs
MenTaLboY:	i.e. differences of 1 blue bit
MenTaLboY:	I can see differences in one red bit myself
graydon:	and if you ever want to do some operation, you will need to know whether colors need to be recomputed, or whether you can use a hw accelerated function to move them in their native (on-screen) colorspace.
MenTaLboY:	plus, 48 bit gives you more precision (and thus less artifacting) when doing a lot of image ops
MenTaLboY:	graydon: GGI does that too, although I havn't looked at the internals of it
jones:	heh - 24 bits ought to be enough to everyone :)
derry:	does big need to be changed to handle 48bbp then? or is there some way to abstract this
MenTaLboY:	jones: I remember someone saying that about 640K of core :)
Loren:	hold the boat a sec???
graydon:	mentalboy: yeah, my understanding if that you use higher bit resolutions when you're doing funny numerical operations which temporarily jump up a couple of orders of magnitude. That's why audio circutry with 16-bit DACs sometimes get coupled with 20-bit DSPs.
Loren:	I know everything in 24+8bpp was Jim's concept, but in *MOST* cases do the apps care???
MenTaLboY:	not usually
MenTaLboY:	the users might, though
graydon:	loren: they care when they have to compute colors manually, or alter colors manually. Hopefully that'll be a rare occurance, but it does happen.
MenTaLboY:	X allows up to 16 bits per color element for a reason
Loren:	If they don't do graphics they probably don't: just gimme a few widgets and I'm happy...
MenTaLboY:	it wouldn't be so bad if designing for 24bpp didn't also mean degredation of 48bpp images
MenTaLboY:	and believe me, there are people who can see the difference
	graydon ok but this is besides the point: is it possible to use a GGI visual in 24+ 8 BPP?
Loren:	Why not do *ALMOST* like X and let the application request what they need... and then Berlin can do unavailible visuals through emulation... no???
MenTaLboY:	graydon: afaik you could write a display target for that
MenTaLboY:	graydon: the visual thingy is almost infinitely extensible
derry:	perhaps only the internal representation of the bitmaps needs to be changed for now. Use the current 32bpp (24 + 8) for now and add 48bpp later
graydon:	mentalboy: OK, and once you do that, the idea is that the visual is self describing, as is any visual which is open on the real-life display hardware, right?
MenTaLboY:	yes.
graydon:	and because of that self-describing, any operation between them is guaranteed to use hw accel when possible, and use an appropriate roputine the rest of the time?
MenTaLboY:	hell, I'm making a display target that uses terminfo like ncurses (for now it actually uses ncurses) to display text modes on pseudottys and serial lines
graydon:	roputine=routine.
MenTaLboY:	graydon: that's the idea
MenTaLboY:	graydon: in practice, I don't think much code has been written along those lines
graydon:	OK, then I find it really pointless to be thinking about using a visual representation which does not comply with this (obviously useful) architecture. We should use 24 + 8, as was always the idea, internally, and let the GGI layer select the optimal way of getting that data to the screen when we're done making it.
jones:	Perhaps this is the way to gain the "famous" hw accel?
graydon:	but we should use a 24 + 8 ggi visual. Some cards will (or do) support that.
Loren:	jones: part of the way maybe...
graydon:	and if they don't they likely support 24, and we can keep the layout the same in the segment which specifies the 8+8+8 color values.
MenTaLboY:	so, anyone feel like writing a 24+8 target for LibGGI?
	Schubert is away: (Auto-Away after 10 mins) [BX-MsgLog On]
graydon:	mentalboy: If it's the next logical step to getting what we want, then I'll do it, but otherwise I'd appreciate someone else with the skills stepping up.
MenTaLboY:	I've already written 16 and 32 bpp text modes
derry:	mentalboy: are you saying dumb biG and just go to LibGGI?
derry:	dumb=dump
graydon:	guys I have to go. I am closing the log. Can someone else open one?
jones:	graydon: remember to op someone please
graydon:	there are no ops on this network. we're al ops. go ahead -- try to change the topic ;)
MenTaLboY:	derry: maybe
graydon:	bye!
MenTaLboY:	cya gray