fpgroups.com: Revised chat room

[Fairlight]

On Wed, Sep 12, 2007 at 05:44:02AM -0400, Brian K. White may or may not have
proven themselves an utter git by pronouncing:
> The server daemon software, like the client software, both are proprietary.

You say that like it's evil.  You can practically hear it spit out as an
epithet.  TeamSpeak is proprietary, and I have absolutely no qualms using
it considering it's the best, IMHO, nor do all the people I know hosting
TS2 servers on linux boxes.  Ventrilo, likewise, is proprietary, and while
I'm not a vent fan (preferring TS2 instead), it's also not wholly bad
software--I just don't like their emphasis.  In fact, while I can think
of a great many text-based open-source IM and chat packages, I cannot for
the life of me remember reference to -one- VOIP open-source project that's
ready for prime time, much less ubiquitous.

> WHat is my option on say, Palm, today?

Let's be realistic, here.  :)  Although...if SOCOM on PSP does VOIP as part
of gameplay chat (as does KillZone: Liberation), I suppose it's possible.
But that unit itself allows for a headset with a mic.  Does a Palm?  We're
talking about device dependance, here.

> * chatzilla, a mozilla/firefox plug-in, It's sometimes very convenient 
> having an irc client built into the browser so that sometimes you don't need 
> a client at all. It also allows for the use of irc:// urls. Available on 

That really makes no sense.  Chatzilla -is- a client.  Therefore if you
have it, you apparently need a client.  You may not need -another- client,
but as a matter of tacit fact, if you need Chatzilla, you needed a client.

> * CGI:IRC, an irc client written as a cgi, thats inconvenient to set up in 
> that you need to configure a web server somewhere, but if you have that, 
> then from the users point of view that is the ultimate in convenience since 
> it doesn't even require installing a plug-in and it's the only way thats 
> really available anywhere, including locked down kiosks.

It's also a bit of a resource hog.  Although less of one than what Jose has
up right now.  The main problem I'm now seeing with Jose's chat is that
it's polling the "who" list every few seconds on any page of his site.
Then when you're on the actual chat page proper, it's polling that -plus-
polling the chat.  All this happens asynchronously, but the timer is so
tightly wound on both, and it's not like a request has no overhead anyway,
that the browser performance for typing in the message field is noticeably
affected, even on a P4-3GHz/2GB system.  CGI:IRC on the other hand uses NPH
and server push, which is far more efficient, as it maintains a connection
to the server, and it's server-push based, as I recall.  It's slightly
buggy in the last version I tried, but overall pretty good.  But it still
takes server resources you'd rather not expend.

> * x-chat, a standard native irc client, ported to linux/freebsd/windows/mac. 
> Not java, ported, or rather, more like written reasonably portably and 
> simply compiled. Supports lots of features not possible in other clients. 
> There are numerous plug-ins written in numerous scripting and compiled 
> languages, for example there is a text-to-speech plug-in.

I don't recall anyone actually liking x-chat from day one.  Everyone I
remember trying it switched away from it for one reason or another.

> * ircii and/or bitchx, plain text irc clients that work over 
> serial/telnet/ssh/console/screen/mscreen etc...

ircII + LiCE, FTW.  :)

Don't get me started on BitchX...I hate that thing with a passion.

> * GAIM, an all-in-one IM client that knows how to connect to several of the 

Boy, you -are- out of the loop compared to normal.  GAIM is no more and
hasn't been for quite a while.  It's been Pidgin for quite some time,
although it supports all the same things and more.

There's also centericq for curses-based of the same stuff--although the
maintainer has a -horrid- C++ codebase that nobody wants/likes to work on,
and support for MSN and AIM keeps breaking and getting repaired very late
in the game compared to others.

You're also omitting Trillian, which predates GAIM by years, but which
I tossed over for Pidgin, as Trillian is only maintained for security
updates anymore.  It wouldn't talk to Yahoo 8 clients for instance, where
Pidgin will.  Pidgin will also do Yahoo file transfers, Trillian will not.
Cerulean seems to have given up on doing anything but taking in money.

You're also overlooking programs like BackRoom, which are Java-based IRC
client applets.  That one (BackRoom, from Webmasters) is arguably the best
I ever used, and it was available back as far as '98, and I believe earlier
since we were in several major versions when we got into it.  In my
opinion, that's the way to go if you're not going to do Chatzilla.  It
required its own special ircd, however, which was not ideal for all
situations.

> Anyone may also write an irc server, and several of those exist too. The 
> usual advantages and disadvantages of open source software apply. No license 
> counts, no cost, no obligatory help, vast seas of voluntary help, only 25% 
> of which is bad, ability to modify to suit, ability to port to other 
> platforms, deploy anywhere and any how you like...

Need for a system on which to host it, not all ISP's of course being
comfortable with the idea...  There's a downside to that as well.

> like java, tcl/tk, smalltalk, etc.. it's probably hard to find any os or 
> language that doesn't sport at least one irc client, no matter how obscure 
> you try to get. It's even possible to write one in filepro, even without the 
> new net commands. By more than one means even. Via user() and netcat. Or 
> maybe even via plain file io commands and named pipes connected to netcat, 
> or connected to telnet. Or plain file io commands and the virtual device 
> nodes provided by the tcp/ip features built into ksh/bash/zsh.

Or you could shell out the cash to use fP-Sockets.  *duck*  There's also
the "trivial" time investment to actually write such a miserable beast.  I
sure as hell wouldn't want to do it under either model.

There's such a thing as using the right technology for the right job.
There's overlap, but not everything should be written in fP, even if it
could theoretically be done.

> The short version of all that is that there is really no end to the reasons 
> why open source versions of whatever_you_want_to_do are more desireable, 
> useful, & more widely convenient than any proprietary solution. But 

I'm as big a fan of open-source as the next guy.  But considering how much
I've been spending on software lately, I have to say that it's far from my
only criterion.  Robustness, flexibility, and featureset all figure in.  I
could have gone with Scribus for free, but didn't like the idea it was in
perpetual beta status, and thus went with Serif PagePlus, which is wholly
proprietary but which kicks absolute ass.

I fail to see why the OSS zealots insist on open source everything.  I'm as
much in favour of free software as the next person, but I'm also in favour
of things being complete, done correctly, and being relatively stable.  Two
out of three isn't good enough, and you're lucky if most OSS projects have
achieved even that level.  It's at that point, when there are no viable
alternatives that aren't a pain in the ass, that I'm more than willing
to go to commercial alternatives.  I've even recently dumped POVray for
Bryce, DAZ|Studio, and Serif ImpactPlus.  POVray is just too much of a
pain, and the modellers for it all suck in one way or another.  Blender
is powerful as hell (as evidenced by the Elephant's Dream movie [you can
find it on YouTube]), but it's SO poorly documented and complex that I
wouldn't want to go near it without a year's training.  GIMP is still
missing things that not only Photoshop has, but Serif PhotoPlus has.  I
tried OpenOffice.org Draw and it was a sad joke compared to Serif DrawPlus.
The OOO Draw wouldn't even deign to draw the object you were moving as you
moved it if you were just nudging it a bit...so you had to wayyyyyy move
it out of whack to get to the point you could see it while moving it to be
able to move it 1 to 2 pixels away from where it started.  It's just plain
sloppy, period.  Try OOO Draw and then try Serif DrawPlus and look at the
difference.  One is not easily usable or full featured, the other is.  I'll
gladly pay Serif for the quality software, thanks, especially at their
reasonable prices for software that's really designed pretty damned well
and priced far under the actual software's feature/class status.

The only time open source seriously figures into a decision for me anymore
is when something isn't offered for multiple platforms--notably the
platform I'm working with.  Or even an older version of the same platform.
There's a TeamSpeak 2 server for glibc 2.1 yet.  I run it.  But there is
a lot of proprietary software that requires 2.2 or higher, which would be
totally unsuitable for running on an old linux system I have here that
I have no reason to upgrade (and it's not exactly upgradeable...Cobalt
Qube3).  I'm not going to put in a brand spankin' new machine to run one or
even two packages, sorry.  So yeah, having the base for something that you
can back-port or at least compile when a matching binary-only dist is not
available is a plus.

Other than that, the whole open-source thing has lost its novelty,
especially after (for me) 18 years, since way before it was OSS or the
OSI existed.  Emacs, screen (back at orst.edu), etc...been doing it for
ages.  The more I look around, the more I see people jumping on the "open
source" bandwagon for either the wrong reasons, or not even knowing what
their reasons are.  Sometimes they do it just on principle of being too
damned cheap to pay for quality software rather than getting stuck with
something that's in a perpetual alpha or beta state, and whose gamma will
possibly never be completed.  Sometimes it's a Very Good Thing[tm] to have
a Responsible Party in the form of a vendor, whom you can actually pressure
to get stuff done and changes made.  You try to pressure any of the open
source crowd, their response is, "You have the source, hack it yourself."
You go to a company as a paying customer, you at least have some pull,
especially in numbers.  If they want to survive, they'll adapt to what's
requested.  It's simple numbers and economics for them.  You have -no-
leverage with open source developers because it's pretty much a hobby for
90% of them.

Open source is not the holy grail, I'm sorry.  It's a good thing, but it's
not a universal panacea.  And I'll openly disagree with anyone that says it
is, including Stallman and -especially- Raymond--two people I consider
ideologists more than anything else, although at least Stallman's
background is a bit more respectable, IMHO.  I can take him half seriously,
even if I disagree with him.

> sometimes it just sounds so meaningless to say "there are many 
> advantages...", it sounds like all the writer probably really means is "It 
> doesn't cost me $35.50" or "If I copy it, which I am going to do anyways, at 
> least if it's open source I won't get arrested." When really those are about 
> the least of the issues.

Actually, my experience has been that those -are- the biggest issues for
most people.  And for those whom it's not, they swing to the other end of
the spectrum where they evangelise it to the point you'd just as soon not
be associated with it.  As with everything, it's a good idea to at least
experiment with moderation--at least in moderation.

mark->
-- 
The latest synth mixdown...
http://media.fairlite.com/Isolation_Voiceless_Cry_Mix.mp3