The notion of using links in documents to point back to Zope source (and possibly vice versa?) may seem obvious, especially with respect to a Web Centric project like Zope, but it presents a bit of a puzzle. On one hand, Zope brings us a new level of dynamic hypertext, with lot's of wonderful extensions, such as the Catalog, to aid searching, etc. On the other hand the urge to deploy this very technology for the purpose of enhancing the "acquisition" (?!) of Zen smacks us right up against the immaturity of the Web itself. How can the links possibly be kept "fresh", especially with a rate of development as rapid as Zope's? Is "freshness" the sole criteria? What of the historical value of archival code, and documentation? Temporal data integrity may not seem mission critical in this particular instance (Zope as software project), but there are plenty of applications where such criteria are vital. Bobo gave us persistence, Zope gives us integration. Are we ready to tackle the nitty gritty of managing non-tabular data, randomly shifting in four dimensions? Now appears a glimpse of the depths that Ted Nelson addressed in the Xanadu project. It may be time to review the sort of solutions that Xanadu actually applied to tracking documents changing over time. Does anyone have a complete collection of "Byte" from the 80's? There was an excellent article about Xanadu, and a fair exposition of the algorithm. This is getting very interesting!
Temporal data integrity may not seem mission critical in this particular instance (Zope as software project), but there are plenty of applications where such criteria are vital.
Mission critical -- no. Important -- yes. Internet RFCs use a simple but effective mechanism for managing temporal integrity. Each RFC has a unique ID and new RFCs that extend or obsolete old ones say so in their introductions (e.g. http://www.pmg.lcs.mit.edu/cgi-bin/rfc/view?number=822). Wouldn't it be great if "The (old) Product API Tutorial" http://www.zope.org/Members/Zen/howto/ProductAPITutorial said in its introduction "This HowTo has been superceded by "How to Create a Boring Product in Python" http://www.zope.org/Members/gtk/Boring/HowTo-Boring. This doesn't happen now, partly because there is no unique ID ("The (old) Product API Tutorial" was recently move from Amos' folder on the old site to Zen's folder on the new site, so any existing links between the two would have been broken) and partly because we have not established a policy and mechanism for doing so. But let's keep evolving (toward Xanadu?). How about enhancing the HowTo methods on www.zope.org (or perhaps on the next PTK-based version) to assign a system-generated unique ID to each HowTo at the time it's created (or cataloged).
Now appears a glimpse of the depths that Ted Nelson addressed in the
Xanadu
project. It may be time to review the sort of solutions that Xanadu actually applied to tracking documents changing over time. Does anyone have a complete collection of "Byte" from the 80's? There was an excellent article about Xanadu, and a fair exposition of the algorithm.
participants (2)
-
Jerry -
Loren Stafford