Previous years' SoC project ideas suggested porting some or all of a library to C++0x. I wanted to get a read on whether or not Boost developers think that there is value in these kinds of projects. I think there are some interesting pros and cons. One aspect of these projects that worries me about these kinds of projects is that they can force programmers to work in the netherworld between two major versions of the language. They may end up porting the libraries to a CPP abstraction layer rather than actually writing C++0x, which may artificially constrain the ported implementations. I would personally see students write native C++0x than trying to write something in between. Maybe not "porting" as much as "migrating". The obvious downside is that native C++0x ports may not be easily integrated into Boost. Thoughts? Andrew Sutton andrew.n.sutton@gmail.com
On Tue, Mar 9, 2010 at 4:42 PM, Andrew Sutton <andrew.n.sutton@gmail.com> wrote:
[snip]
Thoughts?
I'm not very familiar with the rules that a GSoC project should comply with and what a mentor is responsible for (besides what I've read here: https://svn.boost.org/trac/boost/wiki/SoCSubmissionTemplate), but one possible candidate for a C++0x (or rather C++1x) project could be the Mirror reflection library which I happen to be working on. (souceforge homepage: http://sourceforge.net/projects/mirror-lib/) (docs: http://kifri.fri.uniza.sk/~chochlik/mirror-lib/html/) There are several task, on my TODO list, some of which might be suitable for a SoC project: 1) writing several small/medium-sized or one/two more complex examples of usage 2) writing some high level tools using the low level meta-objects similar to the already existing "factory generator" tool 3) making a tool for automatic or semiautomatic registering of the base-level stuff (namespaces, classes, class members, ...) with Mirror. There is an early version of such tool using the XML output of doxygen. 4) "porting" the library to other compilers besides gcc 4.5.x, which is the only compiler I know of implementing all the C++0x features required by Mirror I'm planning to propose them as master's thesis projects at our faculty this year, but in case of interest by the Boost community I could turn some of them into a GSoC project. Best regards, Matus
I'm having a hard time convincing myself that these would be viable projects for Boost. I've never really seen it discussed on the Boost list. Of course I may have missed it in the daily mail flood :) I feel like its too decoupled from the main product right now. Andrew Sutton andrew.n.sutton@gmail.com
participants (2)
-
Andrew Sutton -
Matus Chochlik