At 12:10 AM 12/9/98 +0000, Jim wrote:
The problem is that Documents are not what you think they are. They are miss-named. They should be called "DTML Methods".
There should be another kind of object called "DTML Documents" that are meant to be the source of information, not a method. They should have properties and search their own attribute namespace before their folder's. Then, when a DTML method, like standard_html_footer, was used in a DTML document, and had:
<!--#var bobobase_modification_time-->
the modification time of the DTML document would be inserted.
I propose, by the end of the Zope beta period to:
- Rename Document to "DTML Method",
- Add a new object type, "DTML Document", and
as a bonus:
- Add a new object type, "Structured Text Document", that would use Structured text input and allow DTML tags to be interspursed, possibly using the Python string format or some other less-verbose format than SSI.
I agree to renaming DTML Methods, but I think a different approach should be taken to Documents. I think we should think more generally in terms of a Content object that can have different encodings like XML, plain text, structured text, XML, HTML, and did I mention XML. That way we don't need a separate type of Object for each separate type of content. As a bonus if Content objects work well, Products like Confera could hold their content in standard Content objects, and so it would be much easier for different types of Products to exchange data. -Amos P.S. I still like to idea I had about making the DTML call stack available. Since I imagine DTML Documents will still want to call DTML Documents, even if DTML methods are available.