IDE Minutes 2009 10 22
- Kin Blas, Adobe
- Javier Pedemonte, IBM
- Jon Ferraiolo, IBM
Nested PROPERTIES elements
Jon: I think this topic is done. Kin's explanation provided what I needed. I updated the schema today and added a sentence to the spec today. I still need to review the spec to see if we need more explanation and maybe an example.
Jon: I just sent email. I am proposing that for a widget all PROPERTY elements must have a unique name.
Kin: Totally agree with you. Another reason is how to resolve same property name when used in a template?
Jon: We need different rules for 'name' in widgets vs APIs. With widgets, unique within document for a given element. For APIs, unique within a given class or singleton. The old wording was there (unique within scope of parent) is probably due to having come from the API side.
RESOLUTION: What Jon proposes for 'name' attribute
Jon: I went over Lori's editing changes. As usual, they are all fine.
Jon: One red-colored note. Asks whether we can merge bottom two sections, the section on entityencode() and escapechars() with the one about unrecognized functions. I don't think what Lori is proposing will work if we change it exactly as she proposes. I agree that the spec could be better, but need to fix some other way.
Kin: Current spec is understandable to me, but I'm close to the topic.
Jon: I'll send email to Lori.
Kin: One more issue on substitution chapter. Does it make sense to do property substitution on inline REQUIRE, PRELOAD and POSTLOAD? Those are not widget-instance things. OK for localization substitution. In our implementation, we don't do those substitutions for this reason.
Jon: Makes sense to me.
Kin: Remove the two bullets about inline REQUIRE, PRELOAD and POSTLOAD?
RESOLUTION: Remove the two bullets about inline REQUIRE, PRELOAD and POSTLOAD?
Jon: I reviewed Lori's editorial fixes. Very minor and all looked good to me. No red-colored issues.
Jon: Lori seems to be able to do two small chapters per week. Just two appendices left now, References and Futures. I hope Lori can get through them in the next week. Then, all that is left is to clean up various red-colored notes in spec. Most of these notes are editorial actions that Lori said she would do.
Empty date value
Kin: Before we go to 1.0, there is one more issue we need to resolve. Back to Dates thing. How do we deal with a date property with empty quotes? Doesn't work in all platforms. If you don't choose a date, we need to figure out what to write into the file. Not current date when file was created, but instead current date at runtime.
Jon: Sounds like a worthwhile feature. First thought: new currentdate() function like entityencode().
Kin: I need to think about this.
Jon: After saying that, I now remember that we had similar discussion previously and ended up inventing the 'outputFormat' attribute.
Kin: I'll send an email with examples.