--On Tuesday, January 25, 2000, 6:39 PM -0500 "Jochen F. Rick"
<[EMAIL PROTECTED]> wrote:
>> Comments?
>
> Sounds exactly like I would do it. Since noone has really inquired much
> into making a new file format, that code is written in one way, but you
> saw that it could easily adapted to make your own format if you so chose.
Well, proxies require it. More generally, I can see wanting to use
something other than the file system as a datastore (e.g., MinneStore, or a
SQL db, etc.).
> BTW, everything you write is making it somewhere on my ToDo list in one
> form or another, so you are actually contributing a lot. It's also funny
> how you trip on things that haven't yet been set in stone.
Glad to here it! :) Lemme add that the Swiki I'm using for my class is very
nice. I like it a lot. You did excellent work!
http://beauvoir.phil.unc.edu/phil31/
(If you want to take a peek.)
> The things that are pretty much set in stone are the the
> addresses/templates/actions/settings/pages separation, backwards
> compatibility for the XML page format, and the template format.
Those seem reasonable.
> Formatters, actions, and addresses are getting close to solid. Start-up,
> administration, and security still need some major work.
Hey! Exactly where I'm poking next ;)
> As far as XML goes, all worship XML!!!
Speaking of XML, the XML/XSL support in VisualWorks 5i is pretty durn nice
(been fooling around with that too).
> Actually, I don't care. XML is at
> least public and easily understandable. Initially, it also helped because
> it is easy to edit by hand. If you have something you like better, fine.
> I think XML is a good combination of being easy to understand,
> cross-platform, and parseable. I would assume that Advantive is doing
> something a little more scalable.
Oh, I definitely thing that the default format should be your XML one. It's
quite nice and flexible. After getting proxies working, I might just do a
ReferenceStream format just to have a reference alternative concrete type
available.
Oh, in order to get page signing working, I had to modify a couple of
"user" methods. Code coming your way soon.
One question about formatters: As it stands, is it feasible to get close
tags, etc. into the picture? I think this is espeically important for <p>
(which is darn tricky if you're going to convert from blanks.), and for
(evenutally) generating XHTML.
I notice that the formatters are a little unintelligent when dealing with
say:
--------------
!A heading
some text
-------------
This gets rendered as:
-------------
<h3>A heading</h3>
<br>
Some text<br>
-------------
Now, it seems clear to me that the return is a source code thing to enhance
legibility, and it's not desirable that it come through (especially given
the default behavior of <h3>). But I don't see an obvious way to get at
this with the formatters.
Now, it may be the case that, at this point, one has to go to a real
parser. (Do you have any idea of the "strength" of the formatters? I.e.,
what grammers can they handle?) A second pass might help as well.
Cheers,
Bijan.
