[M3devel] SVN migration, also promotion

Tony Hosking hosking at cs.purdue.edu
Sun Mar 16 16:40:49 CET 2008


For a small community of volunteers (I don't think any of us are  
making any money actually maintaining CM3), I don't think we are doing  
too badly.  We are much smaller than other projects I am involved in  
(jikesrvm.org for example), which make about the same degree of  
forward progress as the CM3 community is doing.  So, I applaud the  
efforts of everyone involved.  The new regression framework is a  
*huge* improvement in preserving, maintaining, and enhancing the  
reliability of CM3 on many more platforms than most other projects  
support.  If someone (I don't have the time) would like to explore  
migration to SVN and trac I would not object (I happily use both these  
systems in other projects), but the biggest piece of work in any CM3  
transition will be to preserve what we have without too much turmoil.   
I am leery of making any move that will fatally disturb the fine work  
everyone is doing.  I am sure Dragiša meant his suggestions to be  
constructive criticism rather than any kind of attack on individuals  
or decisions made over the long history of the CM3 project.  Remember,  
this project has its roots in a system that was started almost 20  
years ago, before anyone knew what open-source or the internet or Web  
really were.  It is a testament to the staying power of the language  
and implementation that we are still up and running and I think as  
healthy as we have ever been!  I know of very few open-source projects  
that can claim the length of history that Modula-3 and CM3 have.  I  
encourage everyone to stay involved, channel their enthusiasm, and  
make progress happen.

Just my 2 cents... :-)

Antony Hosking | Associate Professor | Computer Science | Purdue  
University
305 N. University Street | West Lafayette | IN 47907 | USA
Office +1 765 494 6001 | Mobile +1 765 427 5484



On Mar 16, 2008, at 8:44 AM, Olaf Wagner wrote:

> Quoting Dragiša Durić <dragisha at m3w.org>:
>
>> And, I forgot to mention probable impact on public with more  
>> organized
>> development of CM3. Most people, and I have few friends from C++/D  
>> world
>> who look regularly at Modula-3, don't see our little community as too
>> organized force. OSS world moved on since CVS became widespread.
>> Separated web, no wiki, no tickets. Source history surfable with one
>> tool, web (best looking in lynx, don't even try something else:)  
>> built
>> with other, no real effort to document in depth.
>>
>> Few weeks ago there was mention about possible promotion time. With
>> state of Modula-3 system we are all aware of, it is excellent idea.  
>> Only
>> problem is - other people must collect plethora of information from
>> plethora of sources. 99% of them will stop at first obstacle, and we
>> have a lot of them ready.
>>
>> One example.. Yesterday, I came over Mr. Hosking's commit log on  
>> m3devel
>> archive. All nice - except text went off right side of my screen some
>> two meters. Most people will pass on - one more unreadable mail  
>> archive
>> on the web. I've used CVS to read it, of course. But most people will
>> not know alternatives, or not care - why would they if we don't  
>> care??
>>
>> Similiar communities are spoiling their "followers", but we Modula-3
>> people are "it's all out there, you are not getting it? Pity!".
>
> I think you don't really realize how much effort has been put into
> getting everything into its current state and keeping the quality
> standard as best as possible. This is because many people have
> contributed their time and work and other resources. Elego, for
> example, has contributed its machines, network bandwidth,  
> administration
> and support, and quite a few months of student's work; I myself have
> contributed months of time to get the 5th release of CM3 running first
> on one and then on more and more platforms since we got the sources
> from Critical Mass.
>
> In recent months I have been working on improving both the test
> infrastructure and documentation/presentation of CM3, while trying to
> support all other contributors as best as possible and coordinating
> their achievements.
>
> I don't think there are a `plethora of sources'; as a matter of fact,
> all resources are available via www.opencm3.net and being integrated
> more and more. There is the web presentation, the CVS repository with
> anonymous and ssh access, CVSWeb, mailing lists, tinderbox regression
> testing, trac (still in testing), etc. Soon coming are contributions
> of CM3_IDE (which is a development environment that really integrates
> different kinds of resources, and already running on some systems,
> but not yet being approved for release by the owner) and packages
> for CM3 integration with advanced version control and `project'
> management. All this needs time and a lot of work.
>
> I don't think it will be furthered by switching from CVS to SVN
> or any other tool change.
>
> I'm a bit tired of these discussions, as I don't think they help
> the project on. What is needed is concrete work on documentation
> and infrastructure. I'd also doubt that switching to SVN would
> bring about a group of motivated and competent contributors
> who have just been waiting for the CM3 folks to change their
> version control to participate in the development of Modula-3.
>
> If there really is the wish within the community to switch from
> CVS to SVN, then those interested should build a team, investigate
> the impacts and propose a migration plan. Elego and me for one
> would need new server setup, the actual repository migration (which
> should preserve as much information as possible), setup of repository
> replication to two machines, integration of SVN into the regression
> test framework, setup of SVN clients, update / rewrite of all the
> repository access information and CM guidelines, and a migration
> strategy for all the existing HTTP links in search engines etc.
>
> If I should make an estimation for this project in a commercial
> environment, I'd say it will take 3-6 months and cost a client
> between 40,000 and 80,000 Euros (that's currently 60,000 to 120,000$).
> I'd rather invest these efforts for other goals. If other CM3 users
> think it is worthwhile to do, they should just do it; after all, this
> is an open source project. No tool will be able to bring about more
> organized development of CM3 simply by its use though. It needs
> willing and competent humans to do this.
>
> I should have used the time it took me to write this mail for working
> on CM3 code and infrastructure.
>
> Olaf
> -- 
> Olaf Wagner -- elego Software Solutions GmbH
>               Gustav-Meyer-Allee 25 / Gebäude 12, 13355 Berlin,  
> Germany
> phone: +49 30 23 45 86 96  mobile: +49 177 2345 869  fax: +49 30 23  
> 45 86 95
>   http://www.elegosoft.com | Geschäftsführer: Olaf Wagner | Sitz:  
> Berlin
> Handelregister: Amtsgericht Charlottenburg HRB 77719 | USt-IdNr:  
> DE163214194
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3devel/attachments/20080316/1f8cd686/attachment-0002.html>


More information about the M3devel mailing list