We really can’t speak of kaizen for such a small project that only develops from by small strokes twice a year, but to take into account users’ feedback I’ve started trying new tools to improve Yapsy‘s quality and make interactions with users a little smoother:
- for a few months now, the sources (originally in a mercurial repository on sourceforge) have been maintained in sync with a github repository. This should simplify the process of improvement proposals from the contributors, many of whom seem to like github a lot and will now be able to do pull requests.
- since the latest release, yapsy is also linked to the continuous integration service Travis CI so that each new commit triggers a full run of all unit-tests.
Now for those that may be interested the following are my impressions about these tools.
Recently while working on yapsy, I fell on a few subtleties related to the creation of Python packages.
I use distutils obviously and the generated packages are uploaded on http://pypi.python.org, but I was struggling with the following problems:
- the data files used by my unit-tests were not packaged
- I couldn’t get pypi to understand that one of the source package was for Python2 installations while the other was for Python3 so that only one of them was displayed on the package page, and of course the few users yapsy has kept complaining that the code didn’t work for obvious reasons like importing ‘configparser’ whereas this module is spelled ‘ConfigParser’ in Python2
The solutions to both problems now seem obvious to me, but were strangely difficult to come across on the web.
Another release of yapsy, my little do-it-yourself plugin library, just before the end of the year, and maybe also the end of the world ;)
This year has brought even more ideas and important features to yapsy via user contributions and stackoverflow questions than previously.
The details of all changes can be found in the release note on sourceforge, but I’d like here to give again all my thanks to this year’s main contributors:
- Mathieu Havel
- Mathieu Clabaut
- Mark Fickett
I’m also proud to say that this year, yapsy has been adopted by the following renown project:
- Nikola, the static blog generator that I heard of on planet Python long before being contacted by its author about yapsy
- err chatbot who has had its own FLOSS Weekly episode not so long ago
A more complete list of project using yapsy is available on the online documentation.
Seeing this small piece of code being adopted by other projects in the wild is a great motivation to maintain it and to make sure it is still relevant !
I’ve recently read a book that is now a bit old according to the web standard and especially since it deals with social web technologies.
This book, Programming Collective Intelligence by Toby Segaran was first published in 2007 and explains chapter after chapter each of the main machine learning algorithms that power (even today still) many of the biggest web services like Google, e-Bay and Facebook.
Despite its publicized social web focus, I can’t help thinking of this book as a more general introduction to machine learning algorithms.
Because of that and also because, due to its age and success, this book already has several detailed reviews over the web, I’ll make an attempt at a kind of cross-review by explaining how it could be a perfect companion book to the on-line and free Machine Learning class by Andrew Ng.
Disclaimer: this started essentially as note-to-self listing a few interesting projects to spare me another internet search session.
RSS feeds (and their twin brothers Atom) are ubiquitous over the internet making it possible to easily get a summary of the latest publications of a given website.
Interestingly a huge amount of websites produce this kind of feeds (most blogs obviously but also sites like twitter[en]) and from this point of view the RSS format is quite lively.
But on the consumer side, I’m pretty disappointed with the “offer” in terms of RSS readers. Over the time I’ve tested several well-known desktop readers (liferea, rssowl, thunderbird…) and most of then ended up synchronizing with Google Reader. This one has consequently come to be my main newsreader and it appears to me as clearly dominating the world of internet based newsreader. However such a predominance is not that much a good sign.