[kepler-dev] Re: [SDM-SPA] New distribution of Ptolemy for LLNL users

Matt Jones jones at nceas.ucsb.edu
Wed Mar 3 12:12:49 PST 2004


Hi Bertram,

I think more frequent releases are a good idea.  That said, here's a couple of
caveats and comments.

1) Personally, I think we should start a more formal release system for
Kepler. In particular, I would like to see us label our releases in terms of
their quality (e.g., alpha, beta, ...).  I'd be fine with releasing the current
CVS head as an alpha release, but I think we'd need to do some more formal
testing to call it a beta, and we're certainly not ready for a full release 
because we haven't completed all of the desired features.  So my preference
would be to release it as "kepler-1.0.0alpha1" or something like that.
I'm not sure what the major version number should be given that Ptolemy is
at 3.0.2, and so we should discuss this.

2) To have more frequent releases, we should have a more stringent list of 
features and bugs that we are targeting for each release.  In the past we've 
used Bugzilla for this and it works well.  Each feature or bug gets labeled 
with a target milestone, and when all of the tasks for a milestone are marked as
done then the release is bundled up.  If some tasks are causing the release to 
be delayed, they can be retargeted at a later release in a project discussion.  
To accomodate this, chad has set up a series of target milestones in the
Kepler bugzilla which I have edited today. The milestones currently are:

    1.0.0alpha1
    1.0.0beta1-05.01.2004
    1.0.0beta2-06.01.2004
    1.0.0-06.12.2004
    Post-rel-1.0.0
    Postpone

So, aiming for the first beta release at the begininning of May in time for the Scotland scientific workflows meeting.  We should discuss the timing of one or 
more alpha releases before that. When new bugs are entered in the system, they
default to the "Postpone" milestone.  The developer who gets that bug should
choose an appropriate target milestone for the bug when they accept it.

3) The installer software should work on Windows, Mac, Linux if possible.  We
have used InstallAnywhere successfully in the past.  It works well on all of 
the OS's listed, and is free.  It also allows you to build installers with and
without a JVM included.  It would be nice if building the installer was 
incorporated directly into the ant build script so that "ant dist" or similar 
automatically built the installer apps for all platforms.

I can't make a conference call this week, but Chad may be able to and that 
should be fine in terms of representing SEEK.

Matt

On Sun, 29 Feb 2004, Bertram Ludaescher wrote:

> 
> Chad, Matt, and Ptolemy members:
> 
> Please feel free to chime into this one. It's originated by SDM/SPA
> folks but due to relevance for Kepler, it will be helpful to hear your 
> POV and coordinate on packaging efforts.
> 
> We can also send around a summary of the phone call, and follow up
> through email/another call (since I know some can't make it; in fact,
> I may be one of those..)
> Bertram
> 
> >>>>> "V" == Vouk  <vouk at ncsu.edu> writes:
> V> 
> V> Monday is Ad Hoc
> V> We do not have an 800 number, but I will arrange for an NC State
> V> MeetMe number (it will cost you long distance charges).
> V> mav
> V> 
> V> Bertram Ludaescher wrote:
> V> 
> >>>>>>> "V" == Vouk  <vouk at ncsu.edu> writes:
> >>>>>>> 
> >>>>>>> 
> V> 
> V> David,
> V> Yes. No problem.
> V> Let's do a teleconference on Monday  (say 9am your time)
> V> to nail down the details.
> >> 
> >> Do we have a regular SPA call on Monday or an ad-hoc one?
> >> 
> >> In the case of the latter, and if we talk about packaging, we could
> >> also invite some of our colleagues from the Kepler team at SEEK/UCSB
> >> and of course Ptolemy/UCB, if they are interested and available.
> >> 
> >> We may also decide to have different packaging efforts (e.g., one
> >> based on InstallShield, one based on InstallAnyWhere, and yet another
> >> one based on Web Start or so -- whichever works best).
> >> 
> >> What does everyone think?
> >> 
> >> Bertram
> >> 
> >> 
> V> Transfac is not quite here yet, but will be next week (actually
> V> Monday it may be available already)
> V> 
> V> We have been snowed-out again so things are not moving
> V> as fast as we would like.
> V> 
> V> No Mac distribution yet. It is not difficult to do,
> V> just different (probably a week to work out bugs
> V> the first time it is done)
> V> 
> V> mav
> V> 
> V> 
> V> David Buttler wrote:
> V> 
> >> 
> >> 
> >>>> Hi,
> >>>> 
> >>>> I would like to create a new distribution of Ptolemy for our users 
> >>>> here at LLNL.  From our conversation a couple of weeks ago, it sounded 
> >>>> like NCSU should be able to quickly create a new version using Install 
> >>>> Shield.  The main reasons I would like a new distribution is to cover 
> >>>> the following concerns:
> >>>> 1) Include the new Transfac service
> >>>> 2) Include the new Web Service actors, to provide access to both the 
> >>>> services available at NCSU as well as other bioinformatics web services
> >>>> 
> >>>> Is this possible to do? Have there been any problems with finding a 
> >>>> Transfac replacement?
> >>>> 
> >>>> As a side question, has there been any work yet on creating a Mac 
> >>>> distribution? Do we have any estimates of how difficult it will be or 
> >>>> when a Mac version may be available?
> >>>> 
> >>>> Thanks!
> >>>> Dave
> >>>> 
> >>>> 
> >>>> 
> >> 
> >> 
> >> 
> V> <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
> V> <html>
> V> <head>
> V>   <meta http-equiv="Content-Type" content="text/html;charset=ISO-8859-1">
> V>   <title></title>
> V> </head>
> V> <body text="#000000" bgcolor="#ffffff">
> V> Monday is Ad Hoc<br>
> V> We do not have an 800 number, but I will arrange for an NC State<br>
> V> MeetMe number (it will cost you long distance charges).<br>
> V> mav<br>
> V> <br>
> V> Bertram Ludaescher wrote:<br>
> V> <blockquote type="cite"
> V>  cite="mid16448.43598.582213.984171 at multivac.sdsc.edu">
> V>   <blockquote type="cite">
> V>     <blockquote type="cite">
> V>       <blockquote type="cite">
> V>         <blockquote type="cite">
> V>           <blockquote type="cite">
> V>             <pre wrap="">"V" == Vouk  <a class="moz-txt-link-rfc2396E" href="mailto:vouk at ncsu.edu">&lt;vouk at ncsu.edu&gt;</a> writes:
> V>             </pre>
> V>           </blockquote>
> V>         </blockquote>
> V>       </blockquote>
> V>     </blockquote>
> V>   </blockquote>
> V>   <pre wrap=""><!---->V&gt; 
> V> V&gt; David,
> V> V&gt; Yes. No problem.
> V> V&gt; Let's do a teleconference on Monday  (say 9am your time)
> V> V&gt; to nail down the details.
> V> 
> V> Do we have a regular SPA call on Monday or an ad-hoc one?
> V> 
> V> In the case of the latter, and if we talk about packaging, we could
> V> also invite some of our colleagues from the Kepler team at SEEK/UCSB
> V> and of course Ptolemy/UCB, if they are interested and available.
> V> 
> V> We may also decide to have different packaging efforts (e.g., one
> V> based on InstallShield, one based on InstallAnyWhere, and yet another
> V> one based on Web Start or so -- whichever works best).
> V> 
> V> What does everyone think?
> V> 
> V> Bertram
> V> 
> V> 
> V> V&gt; Transfac is not quite here yet, but will be next week (actually
> V> V&gt; Monday it may be available already)
> V> V&gt; 
> V> V&gt; We have been snowed-out again so things are not moving
> V> V&gt; as fast as we would like.
> V> V&gt; 
> V> V&gt; No Mac distribution yet. It is not difficult to do,
> V> V&gt; just different (probably a week to work out bugs
> V> V&gt; the first time it is done)
> V> V&gt; 
> V> V&gt; mav
> V> V&gt; 
> V> V&gt; 
> V> V&gt; David Buttler wrote:
> V> V&gt; 
> V>   </pre>
> V>   <blockquote type="cite">
> V>     <blockquote type="cite">
> V>       <pre wrap="">Hi,
> V> 
> V> I would like to create a new distribution of Ptolemy for our users 
> V> here at LLNL.  From our conversation a couple of weeks ago, it sounded 
> V> like NCSU should be able to quickly create a new version using Install 
> V> Shield.  The main reasons I would like a new distribution is to cover 
> V> the following concerns:
> V> 1) Include the new Transfac service
> V> 2) Include the new Web Service actors, to provide access to both the 
> V> services available at NCSU as well as other bioinformatics web services
> V> 
> V> Is this possible to do? Have there been any problems with finding a 
> V> Transfac replacement?
> V> 
> V> As a side question, has there been any work yet on creating a Mac 
> V> distribution? Do we have any estimates of how difficult it will be or 
> V> when a Mac version may be available?
> V> 
> V> Thanks!
> V> Dave
> V> 
> V>       </pre>
> V>     </blockquote>
> V>   </blockquote>
> V>   <pre wrap=""><!---->
> V>   </pre>
> V> </blockquote>
> V> </body>
> V> </html>
> _______________________________________________
> kepler-dev mailing list
> kepler-dev at ecoinformatics.org
> http://www.ecoinformatics.org/mailman/listinfo/kepler-dev
> 

-- 
******************************************************************
Matt Jones                                    jones at nceas.ucsb.edu
http://www.nceas.ucsb.edu/    Fax: 425-920-2439   Ph: 907-789-0496
National Center for Ecological Analysis and Synthesis (NCEAS)
******************************************************************




More information about the Kepler-dev mailing list