Steve Ball wrote:
Mitch Amiano wrote:
I've observed a team of developers using XSLT as a build tool to generate C++
code, instead of using the UML tool that had been selected for them (Rhapsody).
They modelled various aspects of the problem domain using XML, and used XSLT
templates to pump out C++.
[...snip...]
How common is this approach? Anyone else have experience long-term applying a
similar approach?
It would appear that many people are using this technique, and not just
for C++ code (as other respondants have indicated).
I'm not sure what the criteria is to proclaim that the
approach is "common", but it certainly gaining popularity.
Since an XSLT stylesheet may produce text output virtually
any source code may be generated - including most documentation
formats (HTML, nroff, etc).
Yes, I'm curious as to if this seemingly obvious use of the technology is
actually obvious to a lot of people, and if it is actually gaining in
popularity or falling off... and whatever reasons people had individually for
making those choices.
Going further, when one combines XSLT with a dynamic, interpreted
scripting language (Tcl, Python, Perl, etc) it is possible to not
only use XSLT to generate code but to evaluate that code on-the-fly
as part of the runtime application. One of my currently projects
is a GUI XML editor that is implemented mostly as XSLT stylesheets,
using Tcl/Tk as "glue" between the stylesheets and the UI.
HTHs,
Steve Ball
For instance, while some developers love such a thing, other test engineers ask "How do you
test such a thing?" If the answer comes back "with great difficulty", moving forward
with an XML strategy suddenly gets a whole lot harder.
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list