Re: MPCS? and source code discussion (semi-long)

Jason Williams (streak@ccwf.cc.utexas.edu)
Tue, 21 Jul 1998 06:20:45 -0500 (CDT)


On Tue, 21 Jul 1998, david morel wrote:
> Question, who knows the scoop of the actual use of
> h.323 with MPCS?

Depends on what you mean by "scoop" :)

> I downloaded the dreaded 3.x to play around with this,
> from the 3.x machine I could see the netmeeting machine
> without troubles, but the 3.x client was not even in my list of
> connected users in netmeeting.

The 3.X client has to use the H.263 codec (and for audio, the G.723
codec). If the 3.X client doesn't use the H.263 codec, they can't be seen
on Netmeeting. The other troublesome fact is Netmeeting users can never
really tell who the CU participants are. The way I got around that is by
forcing the required video codec (pref-vid-codec command) to H.263,
forcing the required audio codec (pref-aud-codec command) to G.723, and
not allowing any lurkers. This means anyone that can show up will show up
to all participants involved. It's STILL a big hassle though because CU
users can chat amongst themselves and H.323 users can chat amongst
themselves (assuming the T.120 server is running), but they can't chat
with CU users. CU users can't chat with Netmeeting users either. There's
no crossover between CU and Netmeeting with chat like there is with video
and audio. Whenever I played with it, I also had to set the MPCS server
up to switch vid based on a time delay rather than the default of audio
switching. What that means is it will switch who is in the vid every X
seconds rather than whoever used audio last. It works, but it's far from
complete. As I understand it, V.CHAT (or something like that), is the
standard for chat that's arising.

> If I connect with 2.x , I can see the connection made, but netmeeting
> says not sharing... and 2.x does NOT see the netmeeting. (that part was only
> half what I expected) I believe the MPCS docs state that one should be
> able to see other h.323 clients, but would have to use the t.120
> server for chat.

2.X users (as far as I know) can't be seen by Netmeeting unless they use
the H.263 codec. I'm not even sure 2.X users can see other H.323 clients.
Scott at White Pine has told me there's a lot of work going into letting
people have the ability to use the 2.X versions with the H.263 codec to
talk to other H.323 clients. I've never seen it done.

> And now in the midst of what is shaping up to be a pretty good
> fight over the ethics morality and technicalities of code, I have
> come close to making a very crude refmarshal-like
> program. I still have a few weeks ahead of me, but I am getting there.

The question I have is...why another Refmarshal program? I know another
guy writing a similar type of interface for my Refmarshal Proxy server for
his refmons. I just don't get it though. Refmarshal does what it does
quite well. It's a slick interface which works nicely, few bugs, few
crashes, etc. If people just don't like paying for it, that's BS. :)
>From what I've heard about the Visual Basic Refmarshal clone, it'll take a
LONG time before it's anywhere near what Refmarshal is. I just don't
understand the desire to recreate the wheel. Perhaps people are picking
that up from White Pine. :) (The GUI changes, the duplication of a
reflector list when mine works great, the duplication of private chat when
Geektalk worked great, etc.).

> (Streak, exploring expect came yesterday, man, this stuff is cool!)

Told ya..expect is nice. :) And it does exactly what you want it to do.

> I am working on it more for my own use then anything else, I will
> probably give it to my refmons as well.. and was considering
> just letting people that wanted to play with it, have it.

Sounds good and all..still dunno why though...I've been trying to promote
the use of my proxy server so the reflector admin can have additions to
Refmarshal to make it "nicer". (logging capabilities..able to restrict
refmon abilites by conference instead of giving every refmon access to the
whole ref...smaller number of REFMON/allow-wpconfig lines to remember,
etc).

Also gotta remember...if you're using expect/TCL, it's an interpreted
language which means anyone else you give it to will have to have expect
and TCL installed and can browse thru your source code.

> Now I am a little nervous about that though. I mean straight up,
> I am "stealing" the idea without question, my whole application
> is based on Dale Paul's refmarshal functionality, and frankly with out the
> first ten lines of code that came from streak in an e-mail, I probably
> would never have had the confidence to learn expect...

As I understand it, that's not illegal. That's kind of like saying
"whoever made CU-OS/2/Q-SeeMe should be sued because they "stole" the idea
of CU and designed an application for it". It's not the same as what
Frank did. I took the "idea" of the Peoplenet reflector scanner and did
it my way..implemented it the way I wanted to, etc. I even added in the
scan command which acted just like the Peoplenet scanner (scanned refs
every X minutes). Does that mean what I did was wrong? no. But if I
purposely ripped off their source code and then advertised it as my
scanner, THAT would be wrong.

> So do I shelf this thing? Am I asking for a brutal beating from the
> most respected names in the cu-scene? Or do I continue on with this?

In one sense, competition is good. But in another, why try to recreate
the wheel when it works perfectly as it is? The "idea" of reflector
monitoring isn't new. Dale Paul had it for about a year before the 2.X
version went public. I had my reflector scanner which offers reflector
monitoring for about a year before the 2.X version went public. The
reflector operator of Fenris has had "CUKiller" out for quite awhile too.
All of these utilities were pooled together. CUKiller was released on the
listserv a year or so ago. Refmarshal improved a lot since the 1.X
versions and was released. My scanner is still used by many for reflector
monitoring/controlling. As far as you're utility goes. Think about why
you're doing it. Dale DOES listen to suggestions (I know I gave him quite
a few..most he implemented..some he didn't).

I know Bill Woodland is creating a nice WP 2.1/MPCS configuration program
called Overlord. In talking with him, he PURPOSELY didn't include stuff
like the "who" commanf to show who's on the reflector. His idea is the
same as mine. Refmarshal works and works great. Why mess with
perfection? Why step on Dale's toes when what he has with Refmarshal is
great. If there's some aspect that you don't like about Refmarshal, email
Dale. I'm sure he could get around to it at some point if it wasn't too
hard to do. The same with my reflector scanner. I've had people email me
suggestions.

> I mean the source, minus the intial log in sequence- which was
> written in proably 11 seconds by streak- is mine. TCL/expect are
> freely distributable, and the solution is only being written for NT to
> interface with MPCS. I guess unix would/should have very little trouble
> running the script though...

TCL/Expect is cross-platform with versions of TCL available on unix,
Win95/NT, and Macs. :)
Also consider the fact that very few are running MPCS. I know Refmarshal
and my scanner (as well as the Refmarshal Proxy Server) are designed to
work with ALL White Pine and Cornell/Enhanced Reflectors. It's a lot of
work maintaining compatibility with all of them. You also never know what
reflector you may end up with. I started out coding for the WP 2.0.1
reflector and had to radically change things with the 2.1 reflector came
out. Luckily, MPCS is almost the same as the 2.1 reflector in its
configuration format.

If you wanted to switch back to the Enhanced Reflector for any reason,
your utility would become useless. (Or if White Pine changed the command
set again in future versions of MPCS).

> But in the end it is about the same as what Frank has done, which is
> take an idea, modify it as he sees fit, and send it back out there..
> [The morality of him using Brian's source is a tricky subject though,
> isn't it? I think I will stay out of that debate]

No...you haven't done what Frank has done. You haven't ripped off the
source code to Refmarshal from Dale Paul, modified it, then distributed it
as your own. You've taken an idea (an idea of "Reflector monitoring"
presumably) and then implemented it. Implementations of ideas have always
been different. White Pine "implements" CU-SeeMe in a much different way
than Cornell "implements" it.

> Am I treading on some intelectual property rights?

I don't think so...but I still wonder what Refmarshal lacks. Your target
market is the same as Refmarshal (ie: overlapping markets) which means
competition. I've been looking at OTHER ideas for Refmarshal. I'd like
my Mac refmons to be able to use Refmarshal so I've been tossing around
the idea of a Java version of it (which would mainly interface with my
proxy..unless I were to do it as a Java application instead of an
applet..yucky security restrictions)

> (Note: I quit comp sci and became a philosophy major last year)

Philosophy? hmm.. You have lots to ponder? :)

> I guess what bothers me is that there are a lot of ego's clashing here
> and in the end we should all be on "the same team"

I don't see it as egos clashing...Frank STOLE Brian's code, changed it
slightly, and turned around and is distributing it as his own original
code. That's illegal, immoral, and unethical. Brian has every right to
protect his own intellectual property. The "same team" doesn't include
ripping people's code off. I'd be pissed as well.

Especially when I'm sure Brian would have answered any of Frank's
questions..I know Brian has been most helpful in helping me understand a
bit more about the way CU works.

The "same team" offers help to its members, but it doesn't steal code from
its members. :) There's a big difference.

> The same team may be just be another way of saying:
> Wpines 3.x is still lousy.

Some things never change. :)

> Seriously, for a lot of us CU is more then a hobby, it is an intrical
> part of our lives.

How true...Some of us use it constantly. :)
If you are addicted to CU-SeeMe, join us on the reflector as we begin the
12 step program. Of course, the meetings are daily on a reflector. :)

> I have devoted a large chunk of the last two+ years in one way or
> another to it, and I am newcomer by most of your standards.

Nahh..not a newbie...I've only been around since late 1993-early 1994. CU
still captures my attention. I've always told myself, "once I get bored
with it or find something better, I'll move on". That day has never come.

> Perhaps I have been lucky, I have been given a lot of help along the
> way.. E-ref and a TON of help from streak the most notable.

What can I say? I enjoy helping people get up and running and
experiencing CU at its finest. Struggling with capture cards, video
formats, reflector config files, etc. can be a real detractor to CU (or
any program).

> But I do have this app on the verge, well if I can get the damn array to
> function properly.. what do you think? Am I asking for trouble with
> something like this or do I finish it up and see what it does?

I wouldn't say you're asking for trouble...If it's an Expect app, it will
ONLY run on Unix and NT (no working Win95 port of Expect that I know of),
so it may be quite limited. If you're using Expect for the whole thing,
it sounds much more like what my Proxy server does than what Refmarshal
does. The Proxy runs as a server..People connect to it and type in
reflector commands to a reflector that the proxy connects to. Maybe not.
:)

> And don't forget about my actual intent in this mail, which was to get
> the scoop on the reality of h.323 :)

Scroll wayyyy back to the top. :)

> I will refrain from sugesting a group hug....

Thank you! :)
Group hugs reminds me too much of what I've seen from AOL. :)

--
streak@ccwf.cc.utexas.edu    * Jason Williams -- Austin, Tx.  |     |
streak@mail.utexas.edu       * University of Texas at Austin  | ___ |
streak@cs.utexas.edu         * BS Computer Science             \_|_/
*************** http://ccwf.cc.utexas.edu/~streak/ **************|