From jean-marc.valin at hermes.usherb.ca  Thu Nov  1 01:19:45 2001
From: jean-marc.valin at hermes.usherb.ca (Jean-Marc Valin)
Date: Fri Feb 10 19:20:54 2006
Subject: [Pipet Devel] IRC meeting times
In-Reply-To: <3BE022F2.33287EE6@bioinformatics.org>
References: <3BD7326F.92B2FAF7@bioinformatics.org> <3BE022F2.33287EE6@bioinformatics.org>
Message-ID: <1004595585.3be0e9818f6b3@imp.usherb.ca>

> That would be 11:00-12:00 EST (daylight savings gives us GMT-500 now). 
> And
> I guess JMV is in GMT-600.
> 
> Any objections?  If not, we'll have our first formal meeting Nov 6.  Of
> course, we can meet informally before then.

Actually, I'm on EST and I'm usually busy from 11:00-12:00 EST. How about
12:00-13:00?

    Jean-Marc


From jeff at bioinformatics.org  Thu Nov  1 13:25:38 2001
From: jeff at bioinformatics.org (J.W. Bizzaro)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] IRC meeting times
References: <3BD7326F.92B2FAF7@bioinformatics.org> <3BE022F2.33287EE6@bioinformatics.org> <1004595585.3be0e9818f6b3@imp.usherb.ca>
Message-ID: <3BE193A2.8309656E@bioinformatics.org>

Jean-Marc Valin wrote:
> 
> Actually, I'm on EST and I'm usually busy from 11:00-12:00 EST. How about
> 12:00-13:00?

That's fine with me.  Then, unless there are objections, we'll meet Tuesdays
at 17:00-18:00 GMT/UTC.

Cheers.
Jeff
-- 
J.W. Bizzaro                                jeff@bioinformatics.org
Director, Bioinformatics.Org        http://bioinformatics.org/~jeff
"As we enjoy great advantages from the inventions of others, we
should be glad of an opportunity to serve others by any invention
of ours; and this we should do freely and generously."
                   -- Benjamin Franklin
--


From jeff at bioinformatics.org  Tue Nov  6 11:04:21 2001
From: jeff at bioinformatics.org (J.W. Bizzaro)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] agenda for 11/6 meeting
Message-ID: <3BE80A05.5EAC3B74@bioinformatics.org>

As a reminder, we're going to meet in an hour on IRC

    irc.openprojects.net:6667
    /join #bioinformatics

I'd like to discuss the following:

  1. New developers
  2. Current project design
  3. 0.0.2 release

Cheers.
Jeff
-- 
J.W. Bizzaro                                jeff@bioinformatics.org
Director, Bioinformatics.org        http://bioinformatics.org/~jeff
"As we enjoy great advantages from the inventions of others, we
should be glad of an opportunity to serve others by any invention
of ours; and this we should do freely and generously."
                   -- Benjamin Franklin
--


From jean-marc.valin at infospace.com  Tue Nov  6 12:01:24 2001
From: jean-marc.valin at infospace.com (Jean-Marc Valin)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] agenda for 11/6 meeting
References: <3BE80A05.5EAC3B74@bioinformatics.org>
Message-ID: <3BE81764.BBB7B381@infospace.com>

"J.W. Bizzaro" a ?crit :
> 
> As a reminder, we're going to meet in an hour on IRC
> 
>     irc.openprojects.net:6667
>     /join #bioinformatics

That's where I am (I think). I see nobody... what's wrong?

	Jean-Marc


From jeff at bioinformatics.org  Tue Nov  6 14:31:31 2001
From: jeff at bioinformatics.org (J.W. Bizzaro)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] transcript from 11/6 meeting
Message-ID: <3BE83A92.352E1972@bioinformatics.org>

We missed you there, Jarl :-)
-------------- next part --------------
--> bizzaro (jeff@129.63.145.127) has joined #pipernet
<bizzaro> Hello
--> jmvalin (jmvalin@205.151.224.90) has joined #pipernet
<jmvalin> Here I am
<bizzaro> Sorry, I guess I did say #bioinformatics
--> byang (byang@affy.bimcore.emory.edu) has joined #pipernet
<-- jmvalin has quit ([x]chat)
<bizzaro> I had a meeting in #bioinformatics yesterday, and I wrote that by habbit
--> jmvalin (jmvalin@205.151.224.90) has joined #pipernet
<bizzaro> Any sign of Jarl?
<jmvalin> Nope... Anyone else missing? Brad?
<byang> I guess so. I study computer science and had been working at biomolecular computing resource at Emory for a few monthes.
<bizzaro> Others _may_ come.  I'll keep an eye on #bioinformatics
<bizzaro> So, anyway, Jean-Marc, you know that Bo will be helping us
<-- byang has quit (zelazny.openprojects.net benford.openprojects.net)
--> byang (byang@affy.bimcore.emory.edu) has joined #pipernet
<jmvalin> So, how's Piper doing? I haven't checked much outside Overflow lately.
<bizzaro> Jarl and I took a break for a while, but we're ready to get back into it
<bizzaro> We're about ready to release 0.0.2, which could mean getting more users
<jmvalin> BTW, the good news is that the company I work for (Infospace, speech solutions dept.) has adopted Overflow for part of its product, so it gives me more time to work on that.
<bizzaro> Super!
<bizzaro> Bo: Have you talked to Jarl yet?
<byang> Not yet. Is he coming today?
<jmvalin> ...and Dominic has started using it to control mobile robots, so he's back at developpement.
<bizzaro> Strange, Jarl is usually on this channel all day
<jmvalin> Maybe he's on #piper or #bioinformatics?
<bizzaro> I've been checking #bioinformatics
<bizzaro> We've never used #piper
<bizzaro> I'll mention some of what Jarl and I discussed in Copenhagen
<bizzaro> Regarding design
<jmvalin> I've used #piper with him a couple times.... (Who's bdumm2?)
<byang> Jean, you are focusing on the DL, right? I have seen lots of conversation between you and Jarl. I am looking to fit myself in the BL. Any advises?
<bizzaro> Brian Dumm
<bizzaro> PL
<jmvalin> Well, I'm mostly on the PL actually, a bit a DL.
<byang> Sorry.
<jmvalin> Brad was doing most of the DL stuff, but he had to switch to other work.
<bizzaro> Jarl and I spoke a bit about the DL, since Brad won't be working on it
<bizzaro> We decided to rewrite it
<bizzaro> Mostly because it is difficult picking up someone's project
<bizzaro> Also because we have some design changes
<jmvalin> The Python part or the C++ part in Overflow?
<bizzaro> The Python part will be rewritten, in C++
<jmvalin> Well, then the work is already done...
<bizzaro> Just as the BL was changed from C to C++, we think all of the Piper core should be one language
<bizzaro> What do you mean the work is done?
<bizzaro> Bo: You know C++ right?
<jmvalin> The UI* classes in Overflow/PL (equivalent of the DL) are written in C++. There may be a couple Piper-specific things to add, but probably not that much.
<bizzaro> I know, a lot of the DL is a Python wrapper for the PL
<byang> I'm catching up with it.
<bizzaro> Bo: I'm catching up too ;-)
<jmvalin> Overflow alone is a bit like Piper without a BL. 
<bizzaro> Brad had written the DL to be a way to tie the UIL to the PL
<byang> I have downloaded the tar ball from cvs. Is it the whole source code? Especially the BL part.
<bizzaro> The Piper tarball has everything but Overflow
<bizzaro> Yes, it has the BL
<byang> b/c I don't see a c++ progame in the bl dir.
<bizzaro> Hmmm.  Did you look in the bl/src directory?
<jmvalin> In Overflow, these classes (UI*) are GUI-independent and allow either a GUI or a batch (text) tool to control the PL.
<bizzaro> Jarl and I want the DL to make a "model" of the network before it is sent to the BL (then on to Overflow)
<jmvalin> (I say UI*: that means the classes UINode, UINetwork, UILink, ... it's an object-oriented representation of a computation network to be executed)
<bizzaro> So, we're looking at making the network in DOM
<jmvalin> "making the network in DOM"?
<bizzaro> Apache.org has one of the best DOM implementations
<bizzaro> Are you familiar with DOM?
<bizzaro> Brad started using it in the DL
<byang> No.
<jmvalin> Not at all familiat...
<jmvalin> familiat -> familiar
<bizzaro> Oh, DOM == Document Object Model
<jmvalin> That's the only knowledge I had about it:-)
<bizzaro> It is a way to represent XML as objects
<jmvalin> ...Well, that's *exactly* what my classes do!
<bizzaro> All implementations come with a way to convert an XML document into objects...
<bizzaro> and then objects into XML
<bizzaro> It's much easier to work with the objects than to manipulate XML text
<bizzaro> jmvalin: We can probably then convert DOM -> Overflow classes easily then
<jmvalin> Well, we already have that:-) I never manipulate XML directly. If I needed to switch to something else than XML, it would be trivial.
<jmvalin> I don't know enough (that is, nothing) about DOM to tell.
<bizzaro> In any case, you won't need to use DOM in Overflow
<bizzaro> The BL will handle the conversion
<bizzaro> So...the PL code that is now in the DL should be moved to the BL as a plugin
<byang> Some idea about how BL handle the conversion?
<jmvalin> The only difference is that there's not always a direct mapping between XML to the UI* classes.
<bizzaro> Remember: We had bypassed the BL before, just to get the UIL -> PL connection working
<bizzaro> Now we need to consider how the BL will work
<jmvalin> "he PL code that is now in the DL"?
<bizzaro> Which is what Jarl and I spoke about in Denmark
<bizzaro> PL code <- UI* classes
<bizzaro> (The UI* classes will be addressed by the BL instead of the DL)
<jmvalin> UI* -> DL. The PL classes are lower level. For instance in Overflow there's PL (Node, Network, ...), DL (UINode, UINetwork, ...), UI (GUINode, GUINetwork, ...)
<bizzaro> IC
<jmvalin> Actually, it's much simpler it the BL and the PL/DL communicate with XML instead of objects.
<bizzaro> byang: The interface between the BL and the PL will be handled by a plugin and a translator
<bizzaro> jmvalin: Jarl and I spoke about whether the communication would be via XML or DOM
<bizzaro> jmvalin: We may be leaning toward DOM over CORBA
<bizzaro> (using CORBA to transport the DOM objects)
<bizzaro> The design is very much in synch with current thinking about how to with XML
<jmvalin> I still don't understand why bother with DOM... mostly since all the functions to exchange XML to/from memory/disk/network is done.
<bizzaro> s/to with/to work with/
<bizzaro> I'm not sure
<bizzaro> We can discuss this with Jarl
<bizzaro> I would think that DOM may have something extra
<bizzaro> Some extra functionality that could be useful
<jmvalin> I guess it's better to discuss that with everyone involved...
<bizzaro> Jarl may not be familiar with the UI* classes either
<bizzaro> The UI* classes have XML read/write built in?
<jmvalin> I think it didn't take Brad too long to be familiar with that part. (he wrote some parts of it, and did some cleanup so the DL could use them easily)
<jmvalin> Yes, an xml file is represented by UIDocument and there's a load() and save() method that can handle either memory or streams (file, network, ...)
<jmvalin> Actually, all the UI* classes understand load and save.
<bizzaro> Great.
<bizzaro> Did Brad implement sending UI* over CORBA?
<bizzaro> Or is it all XML text?
<jmvalin> No idea... What I remember is that the "load from XML" function was implemented after Jarl asked for it, so he surely had something in mind.
<jmvalin> What to you mean by "all XML text"?
<bizzaro> all XML text <- Does the DL communicate with the PL by sending XML text, or is it via CORBA objects?
<jmvalin> I guess it's text, since there's no CORBA in the PL.
<bizzaro> Okay, we'll see what Jarl has to say about all this
<bizzaro> byang: Have you been on this channel (#pipernet before)?
<byang> a few times, but didn't see anybody. i should check here more often.
<jmvalin> BTW, Overflow 0.6.0 will be released soon with a couple interesting features. The most interesting: a demo written by Dominic where you run two Overflow/PL instances sending audio through the network (one reading the MP3, the other playing)
<bizzaro> byang: Please keep checking here for Jarl.
<byang> i will, thanks.
<jmvalin> I think the next step could be to make the same demo work using the BL to handle everything (sending the XML, making the communication).
<bizzaro> jmvalin: It's nice you've got networking now :-)
<byang> what kind of communication is not handle by BL, but should be, do you think?
<bizzaro> Me?
<jmvalin> Actually, adding network didn't require any modification to the base classes, just adding 3-4 new nodes.
<bizzaro> jmvalin: Maybe Piper <-> Overflow communication can be via network?
<jmvalin> I think all communications, either DL stuff (XML?) or the actual objects exchanged between nodes, should be handled by the BL. The BL will take care of authentication/security and other stuff.
<jmvalin> I don't think two PL instance should talk to each other directly.
<jmvalin> (What do you mean by "Piper <-> Overflow communication"?)
<byang> why? isn't that what peer-to-peer supposed to be? maybe stupid question since I havn't been familiar with the whole design.
<-- byang has quit (farmer.openprojects.net benford.openprojects.net)
<bizzaro> Piper <-> Overflow <- BL <-> PL
--> byang (byang@affy.bimcore.emory.edu) has joined #pipernet
<bizzaro> Piper <-> Overflow == BL <-> PL
<jmvalin> Actually, what I was discussing with Jarl a while ago is that One process will include the PL and one part of the BL. The BL part will be responsible for communicating with the other BL's.
<bizzaro> byang: The BL handles all peer<->peer communication.  It is the "broker"
<bizzaro> If we're done, I can post the transcript to the list
<bizzaro> byang: if you have anymore question, please don't hesitate to ask, any of us.
<byang> Some questions. In the Interfaces of Layers, I see some *l2ail, what is the ail?
<byang> I saw somewhere, Jarl mentioned the cq networking. What is it about?
<bizzaro> Interfaces of Layers <- which directory?
<byang> on the mailing list archive: http://bioinformatics.org/pipermail/pipet-devel/2001-February/001292.html
<bizzaro> checking
<bizzaro> Oh yeah, that document was partly written by Nicolas, from the Narval project.
<bizzaro> Narval is an "intelligent agent", and it uses artificial intelligence
<bizzaro> So, "ail" is Nicolas's proposal for an "artificial intelligence layer"
<bizzaro> I suppose the program Narval would be the AIL
<byang> is it already implemented somewhere in piper?
<bizzaro> No, not yet
<bizzaro> Narval has been interested in working with us
<bizzaro> But nothing has materialized yet
<byang> how about the cq networking?
<bizzaro> Where have you seen that?
<byang> several places, e.g. http://bioinformatics.org/pipermail/pipet-devel/2001-February/000060.html
<bizzaro> checking
<bizzaro> cq <- I'm not sure.  It is Jarl's abbreviation for something
<bizzaro> Sorry I can't help
<bizzaro> Send a message to the pipet-devel list about it
<byang> alright, thanks.
<bizzaro> So, are you reading through all of the arhives? :-)
<byang> oh, no.
<bizzaro> Note that the further back in time you go, the less seriously you should take the ideas
<bizzaro> I wouldn'
<bizzaro> t bother with anything before March 2000
<byang> ok
<bizzaro> That's when the 3 projects merged
<bizzaro> Are you familiar with the merger?
<byang> no
<-- bdumm2 has quit (Read error to bdumm2[dhcp31174173.columbus.rr.com]: EOF from client)
<bizzaro> Around March 2000, Jean-Marc posted a message about Overflow to the GNOME news page
<bizzaro> I saw it and commented that it was like my own project, Loci
<bizzaro> Except that it didn't have networking
<bizzaro> Jarl also saw it.  He was working independently on the Generalized Messaging System
<bizzaro> Loci had the best developed GUII
<bizzaro> GMS had the best developed network
<bizzaro> Overflow had the best developed data processing system
<bizzaro> So, we contacted each other and asked "why not make one system, combining the strengths of each?"
<bizzaro> We first started calling it VSH, for "Visual SHell"
<bizzaro> That's where we got the mailing list name "pipet-devel"
<bizzaro> We then considered other names, finally settling in Piper, because of the implication of "flow"
--> bdumm2 (bdumm@dhcp31174173.columbus.rr.com) has joined #pipernet
<bizzaro> And also for the many synonyms
<bizzaro> Ever since we've been trying to get the individual parts to work together
<bizzaro> GMS started out as a C program, and we convinced Jarl to change to C++
<bizzaro> Loci is Python, but we may now be switching to C++ as well
<bizzaro> The glue has been CORBA
<bizzaro> But there has been some debate about that
<bizzaro> Brad and Jarl are familiar with CORBA
<bizzaro> While Jean-Marc and myself are less so
<bizzaro> For now, we'll stick with CORBA
<byang> which is good. I have started learning it and found great interest.
<byang> BTW, it seems to me that Overflow is still kinka standalone. Do I need to download it seperatly to make piper work?
<bizzaro> Yes
<bizzaro> We may package Piper with Overflow, but Jean-Marc wanted Overflow to work independently
<byang> Steve, sysAdmin at Bimcore would help me setting up the system. I hope to experience it for a while and may have more questions.
<byang> Thanks a lot!
<bizzaro> He's going to help you set up Piper on your system?
<byang> right.
<bizzaro> What's the OS?
<byang> We are running solaris 8 enterprise 250. we have almost all prereqs for piper. any suggestions?
<bizzaro> Well, I know that most of our libraries run on Solaris
<bizzaro> But we have never compiled Piper on Solaris
<bizzaro> I'm not sure if Overflow has been
<bizzaro> Overflow needs to be compiled first anyway
<bizzaro> See how that goes
<byang> we will be trying. we got gcc 2.95.3 and python 2.1.1
<bizzaro> That seems okay to me.
<bizzaro> It will be nice to see it compiled
<bizzaro> on Solaris
<bizzaro> Just send a message to the pipet-devel list if there's a problem
<byang> Steve is with me now trying on that...
<bizzaro> Just remember: Overflow first, then Piper
<byang> thanks, we will trying this today and tomorrow.
<jmvalin> (sorry for the delay)
<jmvalin> Overflow does compile on solaris x86 (I use it every day)
<jmvalin> So far, I haven't really tried on Solaris SPARC
<byang> our server is sparc. what else beside the tar ball on sourceforge do I need to install Overflow?
<byang> If necessary, we can set it up on linux.
<byang> just downloaded the Overflow-0.5.1.tar.gz from http://sourceforge.net/project/showfiles.php?group_id=590. right way?
<bizzaro> I think there's a list of dependencies in the README file
<bizzaro> For Piper dependencies, check the INSTALL file in the Piper directory
<bizzaro> You should have GNOME and GTK libraries
<bizzaro> Also GLIB
<bizzaro> omniORB
<bizzaro> It should all be listed there
<byang> Yes, I will let you know if any problems pops up. 
<byang> later today or tomorrow.
<bizzaro> :-)
<bizzaro> Okay, I'll post the transcript to the list
<bizzaro> I'll probably step out in a few minutes
From jeff at bioinformatics.org  Mon Nov 12 00:46:58 2001
From: jeff at bioinformatics.org (J.W. Bizzaro)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] agenda for 11/13 piper meeting
Message-ID: <3BEF6252.4938877@bioinformatics.org>

Greetings, Pipers!

I'm going to be in New York City Monday and Tuesday presenting
Bioinformatics.Org to some students at the City College of New York.  That's
during our scheduled meeting time (17:00 GMT on Tuesday), so I will be unable
to attend.

Please, still attend the meeting if you can.  The agenda should be to get Bo
updated on the BL and what needs to be done.  Also, Jarl and Jean-Marc may
need to discuss UI* libs vs. DOM.  When you are done, someone please post the
transcript (buffer) to the mailing list.

Cheers.
Jeff
-- 
J.W. Bizzaro                                jeff@bioinformatics.org
Director, Bioinformatics.org        http://bioinformatics.org/~jeff
"As we enjoy great advantages from the inventions of others, we
should be glad of an opportunity to serve others by any invention
of ours; and this we should do freely and generously."
                   -- Benjamin Franklin
--


From jeff at bioinformatics.org  Tue Nov 20 11:23:50 2001
From: jeff at bioinformatics.org (J.W. Bizzaro)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] REMINDER: IRC meeting at 17:00 GMT
Message-ID: <3BFA8396.304F33C8@bioinformatics.org>

See you there:

   irc.openprojects.net
   #pipernet

Cheers.
Jeff
-- 
J.W. Bizzaro                                jeff@bioinformatics.org
Director, Bioinformatics.org        http://bioinformatics.org/~jeff
"As we enjoy great advantages from the inventions of others, we
should be glad of an opportunity to serve others by any invention
of ours; and this we should do freely and generously."
                   -- Benjamin Franklin
--


From jeff at bioinformatics.org  Tue Nov 27 10:56:43 2001
From: jeff at bioinformatics.org (J.W. Bizzaro)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] REMINDER: meeting at 17:00 GMT
Message-ID: <3C03B7BB.180133C0@bioinformatics.org>

Pipers,

Just a reminder that we will be meeting on IRC today, same place:

   irc.openprojects.net
   #pipernet

in about an hour.

Jeff
-- 
J.W. Bizzaro                                jeff@bioinformatics.org
Director, Bioinformatics.org        http://bioinformatics.org/~jeff
"As we enjoy great advantages from the inventions of others, we
should be glad of an opportunity to serve others by any invention
of ours; and this we should do freely and generously."
                   -- Benjamin Franklin
--


From jeff at bioinformatics.org  Wed Nov 28 06:34:50 2001
From: jeff at bioinformatics.org (J.W. Bizzaro)
Date: Fri Feb 10 19:20:55 2006
Subject: [Pipet Devel] [Fwd: Installing Piper (fwd)]
Message-ID: <3C04CBDA.740551B6@bioinformatics.org>

This message is form Bo (I hope you don't mind that I posted it).  Jeff
-------------- next part --------------
An embedded message was scrubbed...
From: Bo  Yang <byang@emory.edu>
Subject: Re: Installing Piper (fwd)
Date: Tue, 27 Nov 2001 23:09:21 -0500 (EST)
Size: 2294
Url: http://bioinformatics.org/pipermail/pipet-devel/attachments/20011128/c6d4d17a/attachment.mht