Hi Mateusz,
I wonder if there is any roadmap for next nutch versions? Is there any >> place where I could get information what will be included in the next >> release and when? I know I could search in jira but it does not contain any >> information about releases dates. >> > > Nutch 1.8 has the following roadmap https://issues.apache.org/jira/browse/NUTCH/fixforversion/12324326 Nutch 2.x has the following roadmap https://issues.apache.org/jira/browse/NUTCH/fixforversion/12324325 These however are not static and are driven entirely by when committers are available to push a release. AFAIK, we are on the verge of pushing a release candidate for 1.8 right now. You are correct in stating that no dates are associated with the above, this is intentional and is justified by committers time and availability. > > >> >> But updating the documentation is almost completely skipped. >> > I'm sorry I don't agree here. The wiki documentation has been developed over a period of ten years and every now and then it gets an overhaul. We very much appreciate new contributions and also thrive off of new contributions towards areas which we did not know were out dated. > It is extremely hard to find what features are already included in >> particular versions (eg scoring algorithms, api, gui). >> > Every Nutch release comes bundled with a CHANGES.txt. This file is ALWAYS linked to fromt he release section of our website. You can just look there. if this is not sufficient then please suggest something else and we can get to work on improving it. > Even if some of the features are listed in the documentation, it lacks >> information on how to use them (here brilliant example - >> http://wiki.apache.org/nutch/bin/nutch%20nutchserver). >> New features can be added all the time, but without documentation, no >> one will use them. >> > This IMHO is an isolated case for an aspect of the codebase which has not been in full use as of recent. Thank you for brining this to attention though. If you wish to help make this documentation better then please sign up to Jira and provide me with your UID. I'll add you as a documentation contributor. > >> I understand that time and developers are limited. But if contributors >> provide some list of undocumented things, divided into nutch versions, the >> community could improve the docs. >> > Historically speaking, we have had little complaints about the Nutch documentation. It is far from excellent, I accept that, however the main thing people are looking for is features. In this regard the project is doing pretty well I feel. > Nowadays it is extremely hard to say what is included in which version and >> where to find docs. I think it should be a habit to add the doc page (even >> with limited text) to wiki when adding/refactoring features. >> > IMHO this puts an extra burden on developers and contributors to contribute. We already ask them for good quality, well formatted code and a unit test. Do you really think that for minor issues, documentation is essential as well? > What is more, I think that a wiki should be moved to something modern than >> MoinMoin. Eg. Confliece, which is integrated with Jira. >> >> >> In my experience, Confluence is a very slow product. What benefits inlcuding/over and above integration with Jira would this provide us? Why do we need it to be integrated with Jira? Thanks for reading Lewis

