Re: (FC-Devel) FreeCase (fwd)

[Andrey V Khavryutchenko <akhavr@compchem.kiev.ua>]

>>>>> "JMW" == Jeffrey M Wolfe writes:

JMW> I received this via email and wanted to share it with you.  I am going
JMW> to invite the author to join the list.

Here're my thoughts.

[...]

JMW> 1 Completely free, likely GPL.

or LGPL.

JMW> 1 Versioning and locking for objects, diagrams, source code, and
JMW> external documents.  

Versioning -- yes.  Locking -- not sure.  One of the best CM systems, I've
used -- cvs -- goes ok without locking.

JMW> 1 Extensibility - good tools must be customizable.  

I'd like to restate in the following fasion:

1 Extensibility in the form of documented interfaces, conforming predefined 
standart.

JMW> 4 Security in the repository.  

Hmm...  What should be the security model?  I expected this to be solved on 
OS level, not at ours.

JMW> 1 Reporting. At least +HTML and +rtf.

Definitely SGML.  We've to select a DTD and produce user-level
documentation and reports in it.  Good candidates are DocBook (large) and
TEI Lite (tool support is somewhat problematic)

>> From an implementation perspective, we may be able to pick up some of
>> >these features by integrating existing projects.

JMW> At first, there should be defined different layers, as the core (where
JMW> the repository resides) does not need to know the presentation layer,
JMW> where either the user interface is defined or the printed doc is
JMW> generated.

See the _Interface_ requirement.

[...]

JMW> I could contribute a apache-corba interface module, developed for my
JMW> own studies, which would imply, that some of the presentations could
JMW> be done in a Java capable WebBrowser (argo running as an applet?). (A
JMW> reliable ORB is mico, with a nearly full standard Corba 2.0
JMW> implementation)

JMW> The repository could be accessed via the apache-corba interface to
JMW> the, where all security control over the net can be delegated to SSL.

Well, I'd vote for yours, Jeff, idea of secondary interface servers.

[...]

JMW> First hypothesis (very raw)

JMW> Most things are containers.  Diagrams contain classes, objects,
JMW> relations Classes contain attributes, methods Methods contain
JMW> argumentlist (in , out), a body etc.

I'm sorry to say, but this and the following two paragraphs are a premature
design.

JMW> If all containers are ready, source and basic types can be filled in
JMW> as objects. Therefore, all tools I know, rely on a lisp like
JMW> structure, where containers come from (look into a rational rose
JMW> file!).

JMW> It should be developed a modern container concept, perhaps Java,
JMW> perhaps from libg++, where this can be modeled more easily. I will
JMW> look into argo, if there is a reliable concept and try to map it to
JMW> the needs of the case tool.

If core implementation language will be c++, then we should use STL.

JMW> If this project will succeed, it would case a little revolution in
JMW> developing software.  A few days ago, I have thought about starting
JMW> such a project, however, I do really to have the time to manage such a
JMW> project. Nice coincidence: I either would have called it gnucase or
JMW> freeCase...

:)

Actually, I this the success of this project will be a large revolution in
the field.  That would prove that Net comunity can work not only as hackers
(which is, with all my respect, the Linux kernel team), but as a well
organised and motivated team.

Also, I'm preparing the answer for your post, Jeff, about 'General next
steps'.  Sorry, I've not got much time, but at least I've managed to put
some freecase hours into my schedule :-\

-- 
SY, Andrey V Khavryutchenko	http://www.kbi.kiev.ua/~akhavr

Good judgment comes from experience, and a lot of that comes from
bad judgment.