It seems we need a plan... I will try to invent something:
Prerequisites to deploying solution to wiki.openwrt.org
[ ] All plugins installed in wiki and all configuration done
[ ] tmo26 given enough access to bulk import 900+ txt files
[ ] A few tables with sane filters (Supported, Supported Long, All Devices)
[ ] How to edit entries
[ ] How to create new entries
[ ] Block/lock editing of old ToH to prevent people from editing
[ ] Data values good enough
[ ] Sufficient functionality/performance
[ ] A "go" so tmo26 and I are not thrown out for destroying the wiki ;)
[ ] Things I have not though about???
Where do we stand?
I can imaging two scenarios:
1) We do all the above, and if it does not work out really well, we deal with it.
2) We do all the above, but only as a "test run" with the intention of really doing it in a few weeks. A release candiate. It requires that we
[ ] can delete/overwrite the 900+ data entries at a later point
[ ] we restrict editing of the the data entries, or
[ ] we clearly inform that the data entries will be deleted later, and editing is just for testing
Things that can go wrong?
[ ] Creating/editing entries the new way is not well received
[ ] Entries get edited and corrupted causing unexpected effects
[ ] Performance is unacceptable in production
[ ] Things I have not thought about...
Things that need to be done later:
[ ] Filling in links to openwrt-images
[ ] Filling in other information in dataentries that we currently dont have (mostly links)
[ ] Retiring the old ToH making it clearly obsolete and hard to find by mistake
A few more reflections from my perspective. The Data Entries support quite nice Data Tables and Data Lists. That will be a huge step forward compared to the current situations. However, the Data Tables may turn out to not be "nice enough" (tmo26 mentioned filtering limitations today, for example). Nevertheless, the Data Entries are great, regardless of limitations of Data Tables. It was quite easy for me to harvest information from the current ToH, it will be even easier to harvest information from a Data List in the future. Future opportunities can be
[ ] Improving the Data Tables of DocuWiki
[ ] Creating a Data-JSON plugin to DocuWiki
[ ] Build a JavaScript-based frontend (like the one I have provided), a Data List or Data-JSON as a backend.
I can change my mind if you have good arguments, but I suggest a full test deployment of everything to wiki.openwrt.org... encouraging people to use it and complain about it, but also informing that most likely it is just a test run and edited/added data will be lost.