In the course of updating the Boost Library Incubator website, I gathered some summary statistics which some might find interesting. 40 Number of libraries submitted 2 libraries not accepted into the incubator of lack docs, test, etc. 11 libraries reviews by boost 9 accepted 2 rejected All in all, I'm thinking these numbers aren't too bad for boost and the incubator. I would like to see more submissions - but the submissions we've received seem to be better than average quality and worthy of consideration. Hopefully this has made the review process more efficient. A couple of libraries have also been accepted (and rejected) without having been placed in the incubator. The number is small. I think this validates the incubator as useful, but not necessary and libraries authors should continue to be encouraged, but not required, to submit library candidates to the incubator. Other than the above, I'm not going to try to draw any conclusions from the above admittedly anecdotal data. I would also remind boosters that if you're logged into the incubator, you can display statistics on viewership for a particular library by using the "Display Statistics" button on the library page. This is a crude but someone engaging way to get an idea of how many people have looked at the library page. The attached shows visits to the safe numerics library page in the last 90 days. Robert Ramey
Dear Robert,
In the course of updating the Boost Library Incubator website, I gathered some summary statistics which some might find interesting.
40 Number of libraries submitted 2 libraries not accepted into the incubator of lack docs, test, etc.
11 libraries reviews by boost 9 accepted 2 rejected
I think this was mentioned before in another thread: Shouldn't libraries be removed from the incubator once they are part of boost? See http://blincubator.com/alphabetically/ hana and compute are still listed there. Best regards, Hans
Dear Robert,
In the course of updating the Boost Library Incubator website, I gathered some summary statistics which some might find interesting.
40 Number of libraries submitted 2 libraries not accepted into the incubator of lack docs, test, etc.
11 libraries reviews by boost 9 accepted 2 rejected I think this was mentioned before in another thread: I don't know if it's been mentioned here, but I've been asked about it
On 4/20/17 5:13 AM, Hans Dembinski wrote: directly.
Shouldn't libraries be removed from the incubator once they are part of boost? Why? See
http://blincubator.com/alphabetically/
hana and compute are still listed there. I have changed the "status" setting for accepted libraries from "submitted" to "accepted" Best regards, Hans
-- Robert Ramey www.rrsd.com (805)569-3793
Shouldn't libraries be removed from the incubator once they are part of boost? Why?
The Boost Incubator is called an incubator, so just starting from the name I thought it should only list libraries which are still incubated, not those already hatched and happily running about. There are some practical arguments as well: - Libraries accepted into Boost are listed on boost.org, so listing them also on blincubator seems unnecessary. - Duplicated records which need to be kept in sync manually tend to become inconsistent. - Having a shorter list on the blincubator makes scanning the list of projects by eye easier. Best regards, Hans
On 4/20/17 8:29 AM, Hans Dembinski via Boost wrote:
Shouldn't libraries be removed from the incubator once they are part of boost? Why?
Hmmm - the following are good points and deserve a response. Here it is:
The Boost Incubator is called an incubator, so just starting from the name I thought it should only list libraries which are still incubated, not those already hatched and happily running about. There are some practical arguments as well:
True - maybe I should change the name of the incubator.
- Libraries accepted into Boost are listed on boost.org, so listing them also on blincubator seems unnecessary.
I've always been dissatisfied with the library listing at boost.org. a) It doesn't group libraries like it used to. b) It doesn't have ready access information such as: 1) What version of C++ it supports 2) What the preference is for posting issues 3) Doesn't encourage cloning a single boost sublibrary 4) Other relevant operations. c) I had hoped that the library page would would keep the history of reviews for the library and commentary. And I would like to see more reviews/commentary on the library accessible in a convenient way even after the library is officially accepted. To my disappointment, this hasn't panned out but I haven't given up hope. Perhaps re-implementing the comment/histry mechanism to hook into the boost developer's list would work. Actually, one might think that I have an agenda to replace the boost library page with a better one. He would be correct. It's easier to propose/implement an alternative than trying to get a consensus. Eventually the less useful alternative dies out from disuse.
- Duplicated records which need to be kept in sync manually tend to become inconsistent. Right - it's a pain - for me.
- Having a shorter list on the blincubator makes scanning the list of projects by eye easier. Right - what should do is tweak the script so one could display libraries according to status, functional group, etc. as well as alphabetically by name. Since there's only 40 libraries in there as of now, it hasn't been a priority - and I hate php scripting.
Thanks for your observations.
Best regards, Hans
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
participants (3)
-
degski -
Hans Dembinski -
Robert Ramey