convore.json/groups/python/how-do-you-judge-quality-of-open-source-software/messages.json

[{"user_id": 7, "stars": [{"date_created": 1298394946.9018481, "user_id": 12347}], "topic_id": 8085, "date_created": 1298394770.219214, "message": "No, Yes, Yes, No, No, No :)", "group_id": 292, "id": 180532}, {"user_id": 12416, "stars": [{"date_created": 1298394440.343117, "user_id": 7}, {"date_created": 1298394475.042192, "user_id": 214}, {"date_created": 1298394528.6368771, "user_id": 1289}, {"date_created": 1298398425.8888161, "user_id": 6310}, {"date_created": 1298401036.1134591, "user_id": 927}, {"date_created": 1298402483.2181201, "user_id": 5778}, {"date_created": 1298403818.553292, "user_id": 1}, {"date_created": 1298464117.1646061, "user_id": 6587}], "topic_id": 8085, "date_created": 1298394389.0911479, "message": "You judge open source by reading the source.", "group_id": 292, "id": 180453}, {"user_id": 7, "stars": [{"date_created": 1298724226.850189, "user_id": 11592}], "topic_id": 8085, "date_created": 1298394802.7079239, "message": "code stuff I read mostly, or take on reccomendation from trusted individuals", "group_id": 292, "id": 180538}, {"user_id": 12416, "stars": [], "topic_id": 8085, "date_created": 1298396257.967953, "message": "Those are important but not as important.", "group_id": 292, "id": 180705}, {"user_id": 242, "stars": [], "topic_id": 8085, "date_created": 1298393882.499696, "message": "One of the first things I look for are unit tests.  However I won't necessarily not use software that doesn't have unit tests.  It's also interesting how much consider people (myself included) tend to give to where the source code for some piece of software lives.  But perhaps that isn't actually being used as a quality metric.", "group_id": 292, "id": 180301}, {"user_id": 14243, "stars": [], "topic_id": 8085, "date_created": 1298394955.718775, "message": "well.. I care if the code is horrible.. but if I choose an opensource piece of code, it's because it finished at the point I need... I try not to hack others code until I need", "group_id": 292, "id": 180556}, {"user_id": 242, "stars": [{"date_created": 1298737691.374171, "user_id": 3617}], "topic_id": 8085, "date_created": 1298394744.2967391, "message": "I hope that was an intentionally silly answer.  Because it is a silly answer.   Have you read the entirety of the linux kernel source code?  CPython?  PyPy? Emacs?  Vim?   xnu (for those mac users, yeah it's open source)?  Probably not, but you may consider any or all of these quality software?", "group_id": 292, "id": 180530}, {"user_id": 214, "stars": [], "topic_id": 8085, "date_created": 1298394866.7641721, "message": "Fix a bug, whatever.", "group_id": 292, "id": 180547}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298396692.6359301, "message": "@dreid I can't really imagine someone reading the 3.x 'What's New' documents and coming away thinking 3.x is not better than 2.x.", "group_id": 292, "id": 180756}, {"user_id": 214, "stars": [{"date_created": 1298394853.2431121, "user_id": 7}], "topic_id": 8085, "date_created": 1298394848.605684, "message": "Oh, I care about code quality a lot.", "group_id": 292, "id": 180543}, {"user_id": 12416, "stars": [], "topic_id": 8085, "date_created": 1298395836.269177, "message": "I guess this depends on how you view development and your specific requirements. To me the most important thing to me is: How quickly can I refactor my stack and code in the future? If I haven't read some of the code I'm using I have no data points.", "group_id": 292, "id": 180670}, {"user_id": 242, "stars": [], "topic_id": 8085, "date_created": 1298393948.6804111, "message": "The Zooko Quality Estimator (yeah, that zooko) is an interesting system, the gist is that you look at the last 10 commits to the software, and for any commit that should have a test case and does you give the project 1 point.  You substract a point for every commit that should have a test case and doesn't.", "group_id": 292, "id": 180316}, {"user_id": 14243, "stars": [], "topic_id": 8085, "date_created": 1298394807.8409979, "message": "the idea of opensource is that you use it...", "group_id": 292, "id": 180539}, {"user_id": 242, "stars": [], "topic_id": 8085, "date_created": 1298395464.764261, "message": "@davidhollander I'm not convinced that is an accurate way to judge.  Certainly not in any objective way that would allow you to collect data that would lend itself to statistical analysis.", "group_id": 292, "id": 180627}, {"user_id": 14243, "stars": [], "topic_id": 8085, "date_created": 1298394419.3836119, "message": "lol good one...", "group_id": 292, "id": 180462}, {"user_id": 14243, "stars": [], "topic_id": 8085, "date_created": 1298394828.851198, "message": "i really dont care the quality of it in means of code quality but in product quality", "group_id": 292, "id": 180541}, {"user_id": 214, "stars": [{"date_created": 1298398740.554805, "user_id": 1736}], "topic_id": 8085, "date_created": 1298394859.764642, "message": "If I'm going to use it, chances are I'm eventually going to have to hack it.", "group_id": 292, "id": 180545}, {"user_id": 12416, "stars": [], "topic_id": 8085, "date_created": 1298395037.1892149, "message": "@dreid Use simple random sampling. In statistics it's shown that by randomly selecting samples (X number of functions or classes) you can make accurate estimates about the entire population (every line of code) without measuring each and every one.", "group_id": 292, "id": 180573}, {"user_id": 5639, "stars": [{"date_created": 1298451528.878231, "user_id": 6671}], "topic_id": 8085, "date_created": 1298396396.7229109, "message": "My recent experience porting dateutil to Python 3 has made me put a TON more emphasis on unit tests as a measure of code quality. I always put a pretty high value on unit tests, but even higher now. In a sense they represent a 'summary' of the project's code, assuming the coverage is decent (which also goes into the calculation).", "group_id": 292, "id": 180722}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298396446.800087, "message": "(dateutil has 478 tests, for anyone interested)", "group_id": 292, "id": 180734}, {"user_id": 7, "stars": [], "topic_id": 8085, "date_created": 1298394778.538038, "message": "Depends what I'm using the code for", "group_id": 292, "id": 180535}, {"user_id": 7, "stars": [], "topic_id": 8085, "date_created": 1298394783.4457941, "message": "end user stuff I just use and see if it works", "group_id": 292, "id": 180536}, {"user_id": 14243, "stars": [], "topic_id": 8085, "date_created": 1298394845.268476, "message": "then, if I need to hack it, I may give an opinion about it", "group_id": 292, "id": 180542}, {"user_id": 214, "stars": [], "topic_id": 8085, "date_created": 1298394875.405653, "message": "And bad code quality will make my life miserable at that point.", "group_id": 292, "id": 180548}, {"user_id": 12416, "stars": [], "topic_id": 8085, "date_created": 1298395160.5048971, "message": "aka. just because it would be impossible to read all of the source code for a huge project doesn't mean you shouldn't be reading some. It's the only accurate way to judge.", "group_id": 292, "id": 180606}, {"user_id": 242, "stars": [], "topic_id": 8085, "date_created": 1298396155.336674, "message": "@davidhollander So you base everything on your subjective opinions about the code?  Who wrote it, where they host it, how they've tested it, how they continue to test it, how they've documented it\u2026 none of those things ever come into play?", "group_id": 292, "id": 180702}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298396425.828325, "message": "Going forward, I'll be porting more modules, but I'm far, far less likely to take that on without unit tests to help me validate my work.", "group_id": 292, "id": 180730}, {"user_id": 242, "stars": [], "topic_id": 8085, "date_created": 1298396772.1710019, "message": "@jonesy Eh, I haven't come away with the \"python 3 is so much better than python 2 that I should spend my time porting anything to it.\"  But that is probably best addressed in another topic. :)", "group_id": 292, "id": 180760}, {"user_id": 5639, "stars": [{"date_created": 1298737757.5074999, "user_id": 3617}], "topic_id": 8085, "date_created": 1298396645.8578811, "message": "@dreid but also, Python 3 *is* objectively better than Python 2 if all you look at is the language and not the entire ecosystem, which is lagging in terms of adoption. This is completely expected, but some choose to label Python 3 \"not done\" when in reality it's the various projects that are \"not done\" porting to it. :)", "group_id": 292, "id": 180749}, {"user_id": 5701, "stars": [], "topic_id": 8085, "date_created": 1298396876.757534, "message": "\"Brain surgery\" isn't the usual complaint.  The usual complaint is that at any given time, one can probably think of something else productive to be doing. :)", "group_id": 292, "id": 180769}, {"user_id": 242, "stars": [], "topic_id": 8085, "date_created": 1298396498.412982, "message": "@jonesy Out of curiosity, how did you determine that python 3 was a good use of your time?  Is there some standard by which you believe python 3 is objectively better than python 2.x?", "group_id": 292, "id": 180739}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298396571.5577061, "message": "@dreid I'm working on the 3rd edition of the Python Cookbook, which is only Python 3. Half of the date-related recipes use dateutil. I had the option of spending a total of maybe 3 hours porting dateutil, or days finding/coding/inventing recipes to replace them.", "group_id": 292, "id": 180746}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298396773.5487289, "message": "@dreid between David Beazley and I, probably 5 modules have been ported in the past month (all but 1 by David). All while working full time and working on the book *and* our upcoming Python tutorial at PyCon. It's really not brain surgery", "group_id": 292, "id": 180761}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298396928.823097, "message": "@dreid https://convore.com/python/whats-your-python-3-plan/", "group_id": 292, "id": 180775}, {"user_id": 257, "stars": [], "topic_id": 8085, "date_created": 1298401932.550209, "message": "@jonesy: If most libraries aren't done, and few use it in production, how can you say py3k is \"done\"?", "group_id": 292, "id": 181527}, {"user_id": 257, "stars": [], "topic_id": 8085, "date_created": 1298402138.636476, "message": "e.g. I've ported some code with the help of 2to3, but the resulting code is far from optimal.", "group_id": 292, "id": 181578}, {"user_id": 257, "stars": [], "topic_id": 8085, "date_created": 1298402155.2812021, "message": "as people are now moving, I'd guess there must be many issues showing up", "group_id": 292, "id": 181584}, {"user_id": 5639, "stars": [{"date_created": 1298737787.293901, "user_id": 3617}], "topic_id": 8085, "date_created": 1298404249.9524741, "message": "@asksol you mean 3rd party libraries. A language being done is independent of 3rd party libraries being ported to it.", "group_id": 292, "id": 181978}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298404312.7835259, "message": "@asksol I'd expect some issues as well, but by that metric no version of Python is \"done\".", "group_id": 292, "id": 181992}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298404255.975672, "message": "same for production use.", "group_id": 292, "id": 181979}, {"user_id": 5701, "stars": [{"date_created": 1298451588.786247, "user_id": 6671}], "topic_id": 8085, "date_created": 1298404378.450702, "message": "You can find examples in the stdlib of the not-doneness.  The email package is still a work in progress.", "group_id": 292, "id": 182002}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298404964.6062491, "message": "@jpcalderone good point: I don't really care about it either I guess. My plan is to port things I need to move to 3.x in my production work, but that doesn't necessarily mean I'm going to unleash 3.x in production next week. I'm really trying to be ahead of the curve. If all goes amazingly well, I'll put it in production, given the opportunity. :)", "group_id": 292, "id": 182091}, {"user_id": 5701, "stars": [{"date_created": 1298404850.9435101, "user_id": 5639}], "topic_id": 8085, "date_created": 1298404648.2162409, "message": "(Not that I particularly care about the \"not done\" argument myself)", "group_id": 292, "id": 182052}, {"user_id": 8391, "stars": [{"date_created": 1298424691.472559, "user_id": 1081}, {"date_created": 1298424840.370147, "user_id": 222}, {"date_created": 1298432574.9606049, "user_id": 214}], "topic_id": 8085, "date_created": 1298424652.619561, "message": "Documentation.", "group_id": 292, "id": 185393}, {"user_id": 14923, "stars": [{"date_created": 1298425380.1717279, "user_id": 5639}], "topic_id": 8085, "date_created": 1298422531.4562409, "message": "I generally judge open source software by its documentation.  If the authors have bothered, and the docs are clear and well written, the source is usually pretty solid as well.  I need to figure out if it has a simple-enough-but-not-too-simple interface, anyway.", "group_id": 292, "id": 185319}, {"user_id": 13642, "stars": [], "topic_id": 8085, "date_created": 1298423220.779722, "message": "@jmoiron Agreed.  I get kind of irked when projects rely on tests/specs in lieu of real documentation.  I get that tests serve as the \"spec\" in BDD but real documentation can sometimes be easier to parse and can be organized in a way that makes sense for learning (instead of being organized for code quality, etc.).", "group_id": 292, "id": 185330}, {"user_id": 8391, "stars": [], "topic_id": 8085, "date_created": 1298424679.3683541, "message": "Then tests. But documentation is more important to me than tests. Tests are easy to write. Good documentation isn't.", "group_id": 292, "id": 185397}, {"user_id": 1081, "stars": [], "topic_id": 8085, "date_created": 1298424707.7052701, "message": "@kennethreitz to add to your answer, Documentation & community", "group_id": 292, "id": 185401}, {"user_id": 5639, "stars": [], "topic_id": 8085, "date_created": 1298425459.2016151, "message": "I'd say docs come before tests for me too, but not by a whole lot. Documentation at least affords the hope that the project invites users, collaborators, and eyes on the code.", "group_id": 292, "id": 185437}, {"user_id": 4935, "stars": [], "topic_id": 8085, "date_created": 1298425527.6422801, "message": "commutiny. if there are only joiners, it probably isn't very good.", "group_id": 292, "id": 185443}, {"user_id": 6671, "stars": [], "topic_id": 8085, "date_created": 1298454623.08637, "message": "@kennethreitz I think you're right to consider documentation, but I think it's wrong to say that tests are easier to write.  Writing tests and writing docs are different skills.  Writing tests overlaps a little more with the average programmer's skill set (as it's very much like writing other kinds of code) but it's still different enough that most people get it wrong.", "group_id": 292, "id": 186452}, {"user_id": 6671, "stars": [], "topic_id": 8085, "date_created": 1298454747.5062129, "message": "Personally I judge the quality of a project by the quality of its process.  A really good project is always improving.  A really bad project slides around randomly, some releases improving, some releases introducing tons of new bugs.  Zooko's method seems like a reasonably good quick estimate of this, but I suspect it isn't completely accurate.", "group_id": 292, "id": 186465}, {"user_id": 11592, "stars": [{"date_created": 1298744331.3662319, "user_id": 8391}], "topic_id": 8085, "date_created": 1298725196.6422181, "message": "If project docs aren't done with Sphinx (or just obviously good) it rings a huge bell for me. I'm too tired to dig through epydoc or other crappily generated stuff. Next goes project site - dirty and unelegant appearance more often than not reveals same quality of code.", "group_id": 292, "id": 214004}, {"user_id": 1152, "stars": [{"date_created": 1298744331.5784521, "user_id": 8391}], "topic_id": 8085, "date_created": 1298726774.38714, "message": "Docs, Examples, Tests and finally style and structure and code.", "group_id": 292, "id": 214127}]