Wikibooks:Reading room/Proposals/2018/June

Source Code
There is a persistent problem with source code in Wikibooks. It is very prone to vandalism or introduction of errors, as only a small subset of editors may understand the content and even a smaller one would be inclined to check each edit and the implications to the "correctness" or "compilability" of it. I proposed sometime ago a source code protection for only register editors with the review flag but I guess no one cared about to address the issue. That can take the form of (or a mix of) a Wikimedia software solution or a policy that empowers book communities and reviewers. Another implication to this volatility is also in regard to code licensing (copyrights), but that can only remain a consideration for now and be addressed independently. --Panic (discuss • contribs) 05:40, 2 June 2018 (UTC)
 * Usually the source codes from Category:Subject:Computer programming expose one concept, so they are too small to be compilable or copyrighted. But if someone posts a full 100 lines page of code in one commit, I google that content. Compiling it is pretty quick with https://fiddles.io/. JackPotte (discuss • contribs) 11:43, 2 June 2018 (UTC)
 * It is hard to define what size wise can be copyrightable without a legal challenge, but as natural languages, the rules should be the same, so a simple phrase (ie: routine) can have rights associated. In any case the rights is not the primary concern, as I stated. But I watch several programs (not simple routines) on Wikibooks and several of those have a history of people claiming rights over them (signing into the contribution), I have also watcher rewrites to make a claim and have reverted unreasonable claims. In any case protecting the pages would reduce a lot of wasted work on this area. It would also help enable resume of work, as I'm aware of at least 2 other editors that became frustrated but the code vandalism that happens on the project. I have also stopped contributing code and checking in minor code edits due to this (vandalism/errors) issues (it becomes time consuming to recheck variable changes and non obvious logic). --Panic (discuss • contribs) 12:25, 2 June 2018 (UTC)


 * I understood the question to be partly about protection in the technical (rather than legal) sense. The level of technical protection you're talking about sounds like semi-protection.  However, the platform provides protection per page, and I don't see how anything finer-grained than that would avoid becoming far too complicated to be humanly useable (defeating the purpose of wikis).  Per-page is the level of such things naturally supported by a wiki; which is also inconvenient over at Wikinews, especially when we need to separately review-and-revise multiple pending edits to a published article.  I do not however see anything to be done about it, and I'm rather certain anything of this sort attempted by the Foundation would be catastrophically counterproductive (due to their fundamentally misguided goals). A possibility I have used a couple of times, and have hoped to make much better use of at some point in the future, is some sort of notification category, building perhaps on evalx and even dialog.  The idea is, broadly, that pages check themselves for some expected property (for which evalx is very useful), and if the property is not as it should be, they use a category to flag themselves out for inspection by a human operator.  Then, providing good semi-automation to assist the human operator is where dialog ought to come into play, if I can just get the practical use of dialog to a more sophisticated level.  The one case of this I have in operation on Wikibooks is Category:Attention needed (allbooks), which in the coming upgrade of the subject hierarchy (which I'm still in the process of cautiously designing) looks to be replaced by /Attention needed.  --Pi zero (discuss • contribs) 12:38, 2 June 2018 (UTC)


 * In the only book with code I've made major contributions here, most source code is transcluded (resides in its own page and is included with the text, probably why the problem is made more evident to me as all edits are more clear in function), this helped reuse of examples and maintenance. If for example it is somehow set for reviewer flag level edit protection there is no exclusion of contributions to the "normal content" edits. The change could be in a per book upon request from a reviewer active on that project, even setting some categories for that admin job, this also eases policy change as it can be included on the particular of the reviewer policy. I don't foresee problems as reviewer are invested editors on the project, often project specific and most are contactable.
 * For a patrolled/major editor keeping a large number of these code pages under control is impossible, from indentation changes, format, wikitext for color edits to real code with time a large portion of the code easily becomes corrupted. Page reviews helped at first but at the same time as the number of code changes increased in number or complexity I chose to cease reviewing those edits, sadly to the detriment of readers (since we publish unreviewed edits).
 * The present situation also made me stall a practice textbook (very good to get new editors on the project, and I had planed to link it up with Wikiversity). It seems still to be popular in its trashed state, but I couldn't spend the time reviewing all the re-wrights of already accepted/verified solutions.
 * How complex or easy the solution becomes depends on software limitations and/or admin work, in the end, the reality is that the status quo is not sustainable. Just check the code edits on the algorithms book (and their history for a look on the energy spent keeping it right and often failing) if you have the patience. I also noticed that some books covering less popular languages (or less edited) get less of this error creep (like the ADA book, even if I don't watch each page I haven't seen it on the logs), all others suffer form the situation. --Panic (discuss • contribs) 00:40, 3 June 2018 (UTC)


 * There is no protection level between semi-protect (only autoconfirmed users) and fully protect (only admins). I would be (pleasantly) surprised if the Foundation would allow introduction of a protection level for only reviewers.  We do have the technical ability to set any page so that the most recently sighted revision is visible, as we have done for all the pages at Wikijunior.  Unfortunately, that doesn't help with a transcluded page: even though the page itself displays the most recently sighted revision, the most recent revision is, I believe, transcluded regardless of whether or not it's sighted. There are, however, various opportunities for mild intervention.  In addition to the sort of attention-needed categories I mentioned above, there might be an opportunity to transclude a code page through a template that, in some manner or other, prevents unauthorized changes to the code from appearing prematurely on the transcluding page.  --Pi zero (discuss • contribs) 03:32, 3 June 2018 (UTC)

Redundant language interwiki links
[//en.wikibooks.org/w/index.php?title=French&oldid=3163104&diff=cur Once again today], someone has removed the local interwiki links because they were redundant of Wikidata. Many of them are moreover incomplete, and the other projects editors don't update them, because the English Wikibooks is the only project I know which promotes these good old links since 2015.

The consensus from there wouldn't imagine the frequent differences between for example our links from Statistics and the Wikidata ones, which in addition, are not limited to the other Wikibooks. So I propose to adopt the same system as the other Wikimedia projects to save some time and accuracy. JackPotte (discuss • contribs) 18:01, 5 June 2018 (UTC)
 * Nothing has changed to invalidate any of the reasons I've objected strenuously to mass-destruction of local interwiki information. (In recent months I've noted &mdash;with no bad intentions anywhere in sight, just the error-fostering design of the system&mdash; large-scale corruption of Wikinews interwikis by subtle bot-action on Wikidata.) I do very much want to implement a dialog-based assistant to maintain the relationships between local interwikis and Wikidata.  Having trouble, day to day, conjuring time and energy for the tool development involved.  --Pi zero (discuss • contribs) 18:20, 5 June 2018 (UTC)