|
|
Huge Specs and Good Communication
|
[
Lists Home |
Date Index |
Thread Index
]
In a message dated 11/05/02 01:36:57 GMT Daylight Time, mc@xegesis.org writes:
"The market" most certainly does accept unofficial subsets (oops, sorry
"interoperable profiles"). It even creates new "standards" to rationalize the
profiles, as with XML (a profile of SGML), SOAP (a profile of XML) and what we're
seeing with the WS-I in the web services area.
It's the 750-page "standards" that are generally ignored in the long run.
[actual counter-examples welcomed ...has any "standard" anywhere near as long
as XQuery / XPath2, except for one that exhaustively details existing
practice, been the basis for widespread interoperability ? ]
Mike,
One example of a huge spec that is being well-received and has a rapidly growing developer community is SVG.
The SVG 1.0 REC is 617 pages in the PDF version. The SVG 1.1 CR is 853 pages in the PDF version.
The availability of a PDF version was, in my eyes, a crucial aid to getting up to speed in the (relatively) early days. Navigating a PDF document is so much easier / slicker in my view. The Navigation Pane/Bookmark combination is a huge boost. [Here I could be embarassed by someone explaining an easy way round huge HTML documents that I haven't twigged yet. <grin/> ]
I don't know if its advantages are so perceptible to those (e.g. in a WG) who are already familiar with the material. But for getting up to speed it is excellent.
I would like to suggest to the XQuery/XSLT/XPath WG that they provide a PDF version. Perhaps even a single document for all the XQuery specs with PLEASE a version where all the parts are fully synchronised.
The repeated public WDs which are not fully in synch makes gaining a true overall picture much harder than it need be.
Another crucial factor, in my view, which assisted the growth of SVG is that the Chairman of the SVG WG (Chris Lilley) and the Editor of the SVG Spec (Jon Ferraiolo) were publicly available on an ongoing basis on the main public SVG mailing list to discuss the emerging specification documents. Those guys put in many, many hours to very good effect.
I would like to suggest that the XQuery/XSLT/XPath WG make a couple of individuals with good communication skills available actively to discuss/explain/evangelise in a manner similar to that which the SVG WG did. It will, in my view, pay dividends.
Of course discussions could be "lively" at times, for example when members of the growing SVG community were questioning certain requirements or omissions. By persevering in those discussions the SVG WG got some very useful feedback that will bear fruit in SVG 1.2 and SVG 2.0.
I don't want to appear to suggest that the communication of XQuery is only negative. The clear listing and classification of issues within the various WDs is very helpful as is the change history and color-coded changed material (where supplied). What has been done is useful, but it isn't in my view at least complete.
It appears to have been assumed by the XQuery/XPath/XSLT WG that XPath is (more or less) a subset of XQuery. In a narrow technical sense that may be true. However, the communication of any justification for that assumption/conclusion needs to be improved. If the WG has a good case to make (in much less than 900 pages) then the XPath community will likely be won over. If the WG is making assumptions/presumptions that are not credible/relevant for many in the XPath community then hopefully the combined WG will have the good sense to amend their current direction.
For major technologies a technically sound spec is essential. But it is not sufficient. Excellent, ongoing communication is equally essential.
Andrew Watt
|
|
|
|
|
|