Re: [gmt-dev] UMLX pages and new UMLX paper

["Jorn Bettin" <jorn.bettin@xxxxxxxxxxxxxxxx>]

Hi Ed,

Had a look at your latest paper. Good stuff! My initial comments as follows:

The example is a good practical test for the concrete syntax. No doubt that
the syntax works, it's just quite verbose. If you use a model based template
language you can write that particular example in perhaps 15 lines of
template code... and the result is directly in the concrete syntax of the
target language. So with my practicioner hat on, the UMLX syntax scores
badly. In terms of the concepts used, it scores well, and I agree that we
could map it to whatever the QVT standard is going to be.

Perhaps there is a something to be gained by taking the best of the UMLX
syntax and exchanging some bits against less verbose and more elegant
constructs from template language based approaches. At a basic level your
UMLX syntax is very similar to the less rigorously defined notation I
developed to document model mappings in
http://www.isis.vanderbilt.edu/OOPSLA2K1/Papers/Bettin.pdf and
http://www.isis.vanderbilt.edu/OOPSLA2K1/Presentations/Bettin.ppt. At a more
detailed level there are some interesting differences. The nice thing about
my syntax is the terse notation to specify which elements of the input model
map onto which elements in the output model. If you look closely you see
that I used template-language-like notation for the names of the UML model
elements appearing in the target model - where model element names from the
input model are the template variables.

How about adopting a template language approach to specify the names of
model elements in the target model? This should make the UMLX syntax more
compact and at the same time more readable by removing  the clutter
introduced by having boxes to specify prefixes and names. The result can
still be in abstract syntax, and not concrete syntax.

Furthermore the development of an Open Source model based template language
needs is high on our agenda anyway. This will be the ideal tool to translate
from the abstract syntax of a PSM to the concrete syntax the target
language. The PSM to text transform is also the place where I would
introduce (a) most of the prefixes and suffixes that you may need to
introduce as a result of naming standards etc, and (b) other simple
text-based transformations like removal of blanks from names and
capitalisation of first letters of words etc. Even if you can do these
things higher up the food chain, I'd be inclined to defer them as far as
possible, as this increases the readability of the model and simplifies the
visual transformations.

Regards,
Jorn

Jorn Bettin
jorn.bettin@xxxxxxxxxxxxxxxx
www.softmetaware.com
Tel  +64 9 372 3073 | Mobile +64 27 448 3507 | Fax +64 9 372 3534

----- Original Message -----
From: "Willink, Ed" <Ed.Willink@xxxxxxxxxxxxxxx>
To: "Gmt-Dev (E-mail)" <gmt-dev@xxxxxxxxxxx>
Sent: Thursday, August 07, 2003 8:55 PM
Subject: [gmt-dev] UMLX pages and new UMLX paper


> Hi All
>
> I've finally got round to creating a gmt-home/doc/umlx/index.html,
> with a lonl from parent.
>
> [If some HTML enthusiast cares to improve house-style, I won't be
offended.]
>
> The page links to my latest paper submitted to the M4M workjshop in York
in
> Novemeber.
> I turned my UML to RDBMS example in to a fully fledged paper so it may
> satisfy the
> requirement for a morre convenbtioinally oiriented introduction to UMLX.
It
> also
> has more detail but less surrounding context.
>
> The paper may be found at:
>
>
http://dev.eclipse.org/viewcvs/indextech.cgi/~checkout~/gmt-home/doc/umlx/M4
> M03.pdf
>
> 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)
> _______________________________________________
> gmt-dev mailing list
> gmt-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/gmt-dev