Subject: Re: [boost] Any interest in a framework for memoization?
From: Erik Erlandson (eje_at_[hidden])
Date: 2013-12-14 13:19:52
----- Original Message -----
> > I can only say that I'd like to see library like that in boost but I'm
> > afraid this is not my decision.
> Thank you, Piotr. It appears there is no such interest, at the moment.
Specific to here and now, I assume that the migration to github is occupying a lot of spare cycles in the community.
In my opinion, memoization seems like a useful capability. I have no particular application for it on my to-do list, but having used memoization in prolog before, it seems like the sort of thing I might someday find myself wishing were in boost.
IMO, proposals like this fall into a certain category, that might get less traction because if you randomly selected 1000 programmers, only a handful (at most) would have an application needing the proposed library. I think my current proposal for edit_distance/edit_alignment is also in this category. So is the handful of 'tree' proposals. Most applications simply don't need it Finding the people with the right combination of knowledge/interest in Boost, plus the domain knowledge to give good feedback, is tricky.
And yet... when you want it, you want it :)
I'm not sure what the answer to that quandary is. It's not like boost has none of these. Quaternion/Octonion come to mind as an example.
I've seen other projects (e.g. octopress) that maintain a page of "notable 3rd-party contributions", that aren't in the official project, but can be used as a resource for people interested in capability-X that isn't in the official distribution.
Alternatively, the new github library organization will allow for people to maintain customized library forks, so that is a possible avenue. For example, once the dust clears, I could fork boost::algorithm and maintain my own version that includes edit_distance. Periodically merge in the official version, etc.