[plt-scheme] Sanitizing DrScheme .scm files with XML boxes for MzScheme?

We have been discussion this issue internally for a while.

I have suffered from this XML box thing for as long as they have 
existed. I created an entire Web site with it and then got stuck when 
Matthew changed the format from one version to another and I 
accidentally touched the site with the new version. (I needed the old 
Web server for another week or so.) -- I have suffered from this for 
Interaction boxes in ProfessorJ. -- For test boxes. -- And now I want 
it for the wizard boxes.

So we'll do something soon -- Matthias


On Mar 25, 2004, at 6:21 PM, Anton van Straaten wrote:

>   For list-related administrative tasks:
>   http://list.cs.brown.edu/mailman/listinfo/plt-scheme
>
> Roderick McGuire wrote:
>> Is there someway to sanitize a .scm file written in DrScheme that
>> contains XML boxes so that MzScheme can load it?
>
> I'd like to second this as being a useful capability.
>
> This may not be an original thought, but I haven't seen it discussed 
> much: I
> think "boxes" should really be a view of ordinary textual Scheme code. 
>  You
> should be able to switch between a source view and a "box view" on the 
> View
> menu.  (I can provide a detailed motivation for this, but hopefully 
> it's not
> necessary.)
>
> Ideally, for XML at least, a quasiquoted expression should convert to 
> the
> equivalent set of nested boxes and back.  Of course, some indicator of 
> where
> boxes should appear is presumably needed, but I don't see much wrong 
> with
> something like:
>
>   (define foo (xml-box `(p "woohoo!" ,awesome)))
>
> An operation like 'xml-box' might do nothing in some cases - like the
> above - but for other kinds of embedded content, would translate their
> contents into the corresponding Scheme data.
>
> Also, the above has a strong connection to the question of 
> here-strings,
> which was discussed a few weeks ago.  Matthew said "I'll add 
> [here-strings]
> for v300".  How about tying this into a textual representation for 
> boxes -
> or at least planning for that possibility in future?
>
> Here-strings are typically used to embed another language, which is 
> often
> just text for display, but may also be e.g. SQL code or some other 
> non-sexp
> syntax.
>
> DrScheme now supports boxes for the insertion of various kinds of 
> non-sexp
> content.  A box is essentially a kind of graphical here-string with 
> some
> knowledge of the type of "language" that it contains.  It would be 
> useful to
> extend this concept to support a textual representation, and if this 
> was
> done, the conventional notion of a here-string would be a mere subset 
> of
> boxes.
>
> This kind of thing has some relevance to the commercial world, btw.  
> The
> various web template systems don't tend to have good solutions when it 
> comes
> to IDEs - it's mostly done by text editing, or else by something less
> integrated than DrScheme's boxes.  The ability to graphical edit
> XML/quasiquoted templates, inline, in real code, is pretty impressive. 
>  But
> its use is severely limited if it has no textual representation.
>
> Anton
>