📝 Why #OpenDoc failed, and then failed 3 more times?
Summary of reasons found around the web and comparing them with other implementations of the same concept:
* ActiveX
* KParts
* Bonobo
And a short mention of #WebComponents
Conversation
I really enjoy this topic. A good idea that never quite landed despite a bunch of work going into it. There’s a history here that is fascinating.
1
1
Topic suitable for debate: Is component architecture actually a good idea, given that it never delivered on a fraction of its promises?
2
8
At minimum it's clear that some significant new insight is necessary if the model is to work. Maybe data interchange is really the appropriate layer to target.
1
1
8
The 90s hype-model was "create something so that it that's reusable by others!". Flawed incentive-wise, especially outside big corporations.
Need to drive the other way: "re-use something someone else made earlier!", i.e. demand no buy-in. Hardly any technology takes this line.
2
7
12
But surely re-use is always a function of things being well designed for re-use, with clean interfaces / APIs, loose coupling etc. If the original authors don't support re-use, it's going to be too hard for most would-be re-users to bother.
1
1
1
The problem with interfaces is what's "clean" or "loose" for me probably isn't for you; I can't anticipate that. So need to build tools that apply post-hoc, and bring less cliff-like effort/reward curves. My PhD work was one not-very-successful attempt. cs.kent.ac.uk/people/staff/s
3
1
3
Total aside: as a research-ish systems-builder myself, I really enjoy your bio line: "Mostly I make software slower". I thought you meant: "more deliberately/experimentally"… then I read your site and realized you might really mean "less efficient". I like that both reads work!
Glad you liked it... although it was an accident. :-) I'm fastidious about my non-Germanic, British English approach to adverbs, so only the latter option was intended. I can be pretty slow at most tasks, though.
2