Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
RE: [gmt-dev] Initial CVS structure

Hi All

Glad everyone's listening.

Since everyone responded I cannot respond sensibly to a particular
sub-thread, so I've responded to my original, and since the tabbing
got lost and confused at least Ghica, I've re-typed with 
explicit / paths and put [] around suggestions.
 
gmt-home/doc/umlx				
	(UmlxGraphicalLanguage.pdf) 
org/eclipse/gmt/binaries/NT
	(XmiEmitter.dll, UmlxDecorator.dll)
org/eclipse/gmt/doc/umlx
	(UmlxGraphicalLanguage.doc)
[org/eclipse/gmt/examples]
[org/eclipse/gmt/examples/compiled]
[org/eclipse/gmt/examples/models]
org/eclipse/gmt/libraries
	(nicexsl.jar, saxon.jar)
org/eclipse/gmt/tools
org/eclipse/gmt/tools/gme
org/eclipse/gmt/tools/gme/XmiEmitter
	(~20 DevStudio sources)
org/eclipse/gmt/tools/gme/UmlxDecorator
	(~20 DevStudio sources)
[org/eclipse/gmt/transforms]
[org/eclipse/gmt/transforms/compiled]
[org/eclipse/gmt/transforms/models]
org/eclipse/gmt/umlx
org/eclipse/gmt/umlx/bootstrap
	(~20 NiceXSL,  XSLT sources)
org/eclipse/gmt/umlx/compiled
org/eclipse/gmt/umlx/compiled/xslt
[org/eclipse/gmt/umlx/compiled/java]
org/eclipse/gmt/umlx/language
	(Umlx.mga, Umlx.xmp)
org/eclipse/gmt/umlx/models
	(UmlxCompiler.mga)

Areas
-----

The main areas are:

binaries
  for(re-)distributed executables without sources
  for images that not all developers may be able to re-build
libraries
  for (re-)distributed Java
tools
  for locally developed tools
  for add-ins/ons to third party tools 
umlx ...
  for my UMLX compiler

If we are dealing mainly in transforms as a programming language
then transforms/.../models may be a plausible place for the source
and transforms/.../compiled for the executable form.

Doc
---

> From: "Jeff Hoare" <jhoare@xxxxxxxxxxxxxxxx>
> (a) location of documentation, ie will there be a central 
> location for all
> (ie
> a documentation project if you like), or will each component 
> provide this,
> or
> perhaps we have a combination of both.

I think it is better to do

org/eclipse/gmt/doc/umlx

rather than 

org/eclipse/gmt/umlx/doc

so that the documentation source tree

a) matches the publication tree gmt-home/doc/umlx
b) can be easily distinguished for CVS updates

- with slow serial links I may not want to update all the
bitmaps embedded within large PostScript files.

Build
-----
> From: "Jeff Hoare" <jhoare@xxxxxxxxxxxxxxxx>
> (b) The build structure (which may affect the structure within each
> component).

My current code is all non-Eclipse, so has local builds.
When we are prtoperly integrated with Eclipse, this should
be provided in a consistent Eclipse fashion. I suspect that
Visual C++ builds will always have to be independently
structured. Hopefully we can lose them.

> From: Jorn Bettin> 
> The suggested approach gives everyone a chance to evaluate the initial
> assets that are submitted. The initial submissions from 
> various corners may
> not be compatible amongst each other and may - in parts - be 
> in conflict
> with the intended GMT architecture. Let's discuss the core 
> tool architecture
> in the form of UML diagrams first, before we jump in and 
> carve out elaborate
> directory structures.

Yes. But, I have initial assets that I need to park somewhere. 

umlx is probably temporary since I would expect
to change the name to qvt once the standard emerges, or to gmt 
or compiler if some greater core tool philosophy emerges.

libraries and tools/gme are hardly elaborate directory structures -
merely an attempt to establish some order from the outset.

	Regards
			
		Ed Willink

------------------------------------------------------------------------
E.D.Willink,                             Email: mailto:EdWillink@xxxxxxx
Thales Research and Technology (UK) Ltd, Tel:  +44 118 923 8278 (direct)
Worton Drive,                            or  +44 118 986 8601 (ext 8278)
Worton Grange Business Park,             Fax:  +44 118 923 8399
Reading,   RG2 0SB
ENGLAND          http://www.computing.surrey.ac.uk/personal/pg/E.Willink
------------------------------------------------------------------------
(formerly Racal Research and Thomson-CSF)

As the originator, I grant the addressee permission to divulge
this email and any files transmitted with it to relevant third parties.


Back to the top