Wikipedia:Bots/Requests for approval/Helpful Pixie Bot 48

Trivial.Rich Farmbrough 15:12, 24 March 2012 (UTC)[reply]

Code written,testing. Rich Farmbrough, 19:03, 24 March 2012 (UTC).[reply]

Seems sound. Rich Farmbrough, 19:07, 24 March 2012 (UTC).[reply]

Not really sure what is being asked... do you have an example edit (done manually or by the bot in a sandbox?) Headbomb {talk / contribs / physics / books} 21:02, 24 March 2012 (UTC)[reply]

here. Rich Farmbrough, 22:35, 24 March 2012 (UTC).[reply]

Except there is no reference... adding the section would be misleading and make the article much uglier. Headbomb {talk / contribs / physics / books} 23:50, 24 March 2012 (UTC)[reply]

Yes the point is in the description "Will add a references section, per request, which will be populated when the infobox is changed." AIUI the infobox will generate at least one reference, which would otherwise be left causing an error. Please see Wikipedia:Bot requests/Archive 47#add ref section to language stubs. Rich Farmbrough, 00:44, 25 March 2012 (UTC).[reply]

Some of the changes in the test edits violate the bot operator's edit restriction by making cosmetic changes that AWB does not make, such as:

• Changing the capitalization of the first letter of templates [1] [2] [3]
• Replacing <references/> with {{reflist}} [4]. This is particularly strange because the task refers to missing ref sections, so if there is a ref section then it does not need to be changed for this task.

These should be turned off before the task is approved; there is no need for the bot to make such changes in order to accomplish the task that as described. — Carl (CBM · talk) 23:49, 24 March 2012 (UTC)[reply]

WP:COSMETICBOT only says that such edits should not be performed on their own, there is no restriction on bundling cosmetic changes alongside non-cosmetic changes (although capitalizing templates is frowned upon, especially with stubs and citation templates, so that should be turned off). Headbomb {talk / contribs / physics / books} 23:54, 24 March 2012 (UTC)[reply]

I am not referring to cosmetic changes in general; those made by AWB, for example, are fine. However, this testing edit [5] had no effect on the rendered page at all. Since there was already a reference section, the task should not be run at all on that page. I agree about the capitalization, I think that interwikis are also typically not capitalized. — Carl (CBM · talk) 23:58, 24 March 2012 (UTC)[reply]

The former is not a red herring, it is a significant issue with the test edits for this task, and I hope that the BAG reviewers will take it under appropriate consideration. — Carl (CBM · talk) 00:55, 25 March 2012 (UTC)[reply]

They have not been submitted as test edits for this task. This is the sample edit. Now, shhh, I'm working. Rich Farmbrough, 01:19, 25 March 2012 (UTC).[reply]

Just to be clear, that means that the actual bot will not change the capitalization of first letters of templates, nor edit pages that already have reference sections that display footnotes? — Carl (CBM · talk) 01:21, 25 March 2012 (UTC)[reply]

Let's be clear here, while the various concerns raised by CBM are unrelated to this task, they are valid concerns in general and the bot (and you, per that editing restriction) need to abibe by WP:COSMETICBOT. One cosmetic change here and there is not the end of the world, and you should not be crucified for the once-in-a-bluemoon failure, but Helpful Pixie Bot is not exempt from WP:COSMETICBOT policy, and can/should be considered as malfunctioning / editing outside of approval if its failures are systematic. Headbomb {talk / contribs / physics / books} 01:28, 25 March 2012 (UTC)[reply]

That's as maybe. But they are unrelated to this task. I don't go to every place CBM posts and insert insidious innuendo. This is a pathalogical behaviour and needs to stop. Rich Farmbrough, 04:27, 25 March 2012 (UTC).[reply]

Now concerning this task, is there any evidence that the community thinks it's a good idea to add empty reference sections to articles? Because this mind could easily be applied to all stubs, and I highly doubt there is consensus for adding them. Headbomb {talk / contribs / physics / books} 01:30, 25 March 2012 (UTC)[reply]

There is a bot request [6] related to this. They will be changing "infobox language" to include a reference, or adding a reference to all the instances, but the articles need to have the ability to display footnotes to avoid the big red error message when the infobox is changed. — Carl (CBM · talk) 01:33, 25 March 2012 (UTC)[reply]

Adding the section when there is a ref is fine, but not as a placeholder. Headbomb {talk / contribs / physics / books} 01:37, 25 March 2012 (UTC)[reply]

So we let the error happen first, then fix it? Not a problem, I already have approval for that. If you wish to restrict this BRFA to merely adding the new fields,we can get it done before sunrise. Rich Farmbrough, 04:27, 25 March 2012 (UTC).[reply]

I wonder what the fuss is here.

• No, we're not capitalizing the template, we're decapitalizing it, to bypass the redirect. But not on its own as a trivial cosmetic change, only when the template is already being changed anyway. Like ticking it as 'minor' in AWB and then ignoring minor changes in preparse mode, which I do already. Similarly with <references/> vs. {{reflist}}: I'm only asking that we add a ref section if one doesn't exist, as our bots currently do.
• As for the *temporarily* empty ref sections, I've done that to dozens of articles (as when the single ref in an article is bad), and no-one has ever complained. I doubt anyone cares, any more than they care about other empty sections that litter our articles.
• No, this wouldn't apply to other stubs, because other stubs don't have templates that auto-generate reference footnotes. All of these articles should be overtly referenced, rather than covertly referenced as they are now. People have complained about this, which is why I added the auto-ref capability to the template in the first place. Newbies even come by and tag them as 'unreferenced', because they don't notice the covert ref, and I follow them around and revert them. However, we can't add the actual refs to all of the articles right now, because we need to go through manually and verify that the ref reflects the current state of the article. (It does 90% of the time, but that other 10% is several hundred articles.) I hope to get WProj Languages together to review them (I suggested this some months ago): if everyone took 100 articles, it wouldn't take long, and with this bot request, it would be even easier and the refs would be properly formatted.
• Sure, we could manually generate thousands of errors for you guys to clean up if you like instead. I take this to mean there is no reason for me to clean up an article when I edit it, that I should just leave a mess and let someone else take care of it? Because up till now, I have sometimes not ref'd an article because I was tired and didn't feel like typing up yet another ref section. Which is silly, but I hate leaving a mess.
• This[7] is very strange. I can see not approving a bot, but wasting processing power reverting trivial changes that cause no harm? That would seem to violate the spirit of the very WP:RESTRICT guideline that's being used to justify it. Again, this is no different that marking trivial/cosmetic changes 'minor' and skipping them when scanning with AWB. Why would anyone go around and revert cosmetic changes when they are made as part of a larger edit?

— kwami (talk) 05:30, 25 March 2012 (UTC)[reply]

This edit [8] capitalized the first letter of "ie-language-stub". That does not bypass a redirect, or have any other effect on the rendered page, and in fact it goes against common practice. As for the motivation in reverting some of the changes, the bot operator has a long history of making these undesired cosmetic changes, to the point that he has an actual edit restriction against making them. So these edits are not supposed to have been made in the first place. — Carl (CBM · talk) 19:26, 25 March 2012 (UTC)[reply]

Regarding adding new fields to the infobox: are values actually being added for the fields, or are they just being inserted empty? Since the latter has no effect on the actual parsing of the template (empty parameters are the same as none at all), I don't see why the bot needs to insert them, particularly if it is not doing anything more significant at the same time. The same goes for rearranging the order of parameters: the parser does not care what order they are in. — Carl (CBM · talk) 19:26, 25 March 2012 (UTC)[reply]

Once again you miss the most fundamental point, the most fundamental point - that we are human oriented. If inserting empty parameters makes it easier for editors it is a good thing. If removing empty parameters makes it easier for editors that is a good thing. If re-ordering them makes it easier that is a good thing.

Now that is out of the way please note that an empty parameter is not the same as none at all.

Also please read the request, it is only asking for the extra fields when the reflist is being added, although adding them in all cases would be totally reasonable. Trying to second guess WikiProjects on detail like that (as opposed to fundamentally flawed policies such as you steamroller in WP Maths) is pure lunacy.

Rich Farmbrough, 20:27, 25 March 2012 (UTC).[reply]

CMB, these are all fields we *should* have in a language box: we should have a name, SIL code, classification, region, population figure (unless extinct), date, reference, and – if one exists – a link to our ethnography article. All the other params are optional. But the fundamental ones – if they do not exist, an editor should add them in. Adding them in is quicker and easier if the entry already exists. Also, it's easier when they are in the expected order. I've found numerous cases were a param has been added twice because of ordering problems. I can go through and do this all with AWB, but it would take forever. Much easier with a bot. — kwami (talk) 21:14, 25 March 2012 (UTC)[reply]

I also note that the bot operator continues to make edits that appear to be testing edits for this task (e.g. [9] [10]). Perhaps BAG could ask him to stop until the task is actually approved? — Carl (CBM · talk) 19:26, 25 March 2012 (UTC)[reply]

Good. Please indent to your departmental stationary office for a suitable notebook where you can put these notes, and keep them to yourself. Add your matters of record and any other turbidifying snippets you feel inclined to share. Rich Farmbrough, 20:27, 25 March 2012 (UTC).[reply]

Rich, for the last time, stop those edits. You have an editing restriction, and they are not approved as a bot task. And this, combined with your general condescension (see Wikipedia:BOTPOL#Good_communication WP:CIVIL) makes me wonder if we shouldn't revoked approval of all your previously-approved tasks. Headbomb {talk / contribs / physics / books} 20:38, 25 March 2012 (UTC)[reply]

+1. Rich, the edits Carl cites directly violate your editing restriction, which I will enforce via block if these actions continue. While adding a references section where none exists is of course not a cosmetic change, the other changes you are making in those edits do not affect the appearance of the page and do not have BAG approval. I would also strongly encourage you to treat this as a more serious matter and act more professionally as per Headbomb. Hersfold ^(t/a/c) 21:09, 25 March 2012 (UTC)[reply]

Have you no curiosity as to why a professional mathematician deems it his duty to attempt to scupper my BRFAs by bringing up irrelevancies? Perhaps it is the reason that a significant amount of his edits on Wikipedia are edit warring with various editors over trivia. The fact that I refuse to bow to his incessant bullying I would have thought is a virtue rather than a vice. And I would have thought that the fact that I had tried to divert attention back to the BRFA was also a virtue. But when push comes to shove, this is sabotaging activity, if CBM has nothing to add to the substance of the BRFA he should absent himself forthwith. And while I would be polite with an editor who was simply raising technical points he didn't understand, mix that with personal attacks and other passive-aggressive behaviour, and it's time to stop pretending these are good faith objections - either bad faith is present or a level of inability to interact properly with other editors. Moreover playing games such as cycling through half a dozen reasons to oppose something or somebody, and when they are all refuted, starting back with the first, just waste everyone's time, although they may make good propaganda. Rich Farmbrough, 21:16, 25 March 2012 (UTC).[reply]

Now. Do we have any outstanding issues with this BRFA? All this noise is taking valuable programming time. Rich Farmbrough, 21:16, 25 March 2012 (UTC).[reply]

And your refusal to comply with WP:BOTPOL and WP:CIVIL as is expected of all bot operators, and your editing restrictions, is taking valuable community (both BAG and non-BAG) time. I don't care if you like or dislike CBM, but you are not exempt from WP:BOTPOL. So there are two questions here

1) Will you stop editing outside of your editting restrictions, and start editing within WP:BOTPOL and specially WP:COSMETICBOT?

2) Will this bot only add the missing reference section only in the case where it's actually missing, and not as a placeholder?

Headbomb {talk / contribs / physics / books} 21:21, 25 March 2012 (UTC)[reply]

Headbomb, did you not read the request? You don't seem to understand it. If Rich is restricted from making the changes we need, then perhaps someone else should run the bot. Our articles *should* be referenced. That's basic policy. Covert refs cover that, but are inaccessible, and several people have complained about them. We need overt refs. Anything that makes adding them a easier is a good thing. I want the infrastructure set up so that all an editor has to do is check the covert ref, verify that it supports the population figure and is dated properly, and then add "e16" to the ref field. That is quick and easy enough that the languages project can get together and verify our 6,000 articles reasonably efficiently. Having to add the same template infrastructure by hand is a waste of time – this is exactly what we have bots for! These are functional changes. — kwami (talk) 21:29, 25 March 2012 (UTC)[reply]

Not all requests are good bot tasks. Articles should be referenced yes. There's no problem with bots adding references, but if there is no references, it stands to reason that there should not be a reference section in the article. Headbomb {talk / contribs / physics / books} 21:32, 25 March 2012 (UTC)[reply]

Please read the request so that you understand what this is about!

Also, what about adding the infrastructure for necessary parameters, so that it's easier for editors to supply the information? — kwami (talk) 21:46, 25 March 2012 (UTC)[reply]

I did read the request, and I still don't see how how adding an empty reference sections serves any purpose. Adding a |ref= could be done on its own, but there would need to be a clear consensus for that (e.g a discussion on a relevant Wikiproject). Also, if there's consensus for adding |ref= to the infobox, it might also be a good opportunity to do some standardization concerning the infobox (parameter order, add other missing but commonly used parameters, etc...), rather than only add |ref=. Headbomb {talk / contribs / physics / books} 21:53, 25 March 2012 (UTC)[reply]

Which it is doing. Kwami understands the inofbox and how it is used, as I have only created a handful of language articles, I defer to his opinion (hint: I rather think others should too), that these three fields (or less if there is an "extinct" field) are what is needed. Maybe we should fix up the complete array of articles, rather than just those missing a ref section. Those with a ref section are more likely to have the full compliment of fields anyway. Rich Farmbrough, 22:03, 25 March 2012 (UTC).[reply]

@ Headbomb: Yes, actually, I only suggested ordering the fields involved in my request, but if s.o. would like to reorder *all* the fields, so that they match the documentation at {{infobox language}}, that would actually be quite nice. I've already gone through and fixed some of the more egregious shuffled fields (while doing other things with AWB), but I still come across them from time to time. — kwami (talk) 04:35, 26 March 2012 (UTC)[reply]

I figured this would be so trivial that it wouldn't need discussion, but I posted requests here at the wiki project and here at the template. — kwami (talk) 04:48, 26 March 2012 (UTC)[reply]

Speaking of cosmetic changes:[11] — kwami (talk) 05:20, 26 March 2012 (UTC)[reply]

• Comment I found this discussion through a post on Wikipedia talk:WikiProject Languages, and after reading the whole page it seemed to me that had WProject Languages used the time and effort spent programming the bot and discussing here to just do it manually, it would have probably been done by now :) All levity aside though, I see no problem here as long as there won't be empty reference sections left on articles, but (if I understand Kwami's explanations) the reference sections created by the bot would be filled by citations from the infobox. Also since I work mainly on language articles and do other wikignomish things, let me know if there's anything I can do to help out this referencing process. Cheers.--William Thweatt ^Talk | ^Contribs 05:59, 26 March 2012 (UTC)[reply]

The ref field will take any ref, of course, but it's automated for Ethn. 15 and 16, since that's the only source we have for 80%+ of our articles. (Also, it's a named ref, so you can cross-ref it in the text w the tag 'e16'.) The ref section will be blank until s.o. adds the ref, but that's a job for the project as a whole to handle, since there are 3,000+ relevant articles. — kwami (talk) 06:51, 26 March 2012 (UTC)[reply]

I had previously considered using Eth / sil to create language articles, and there's no reason that the referencing can't be automate. Indeed, as William says it could have all been done by now. Rich Farmbrough, 07:26, 26 March 2012 (UTC).[reply]

1. I agree that there should be no empty References section. Now if I understand it well: what is the mechanism (logic) that adds a ref in 80% of the pages? If we apply that same logic in the bot for adding the section, no empty section will be added. (Assuming the reference will be coded in the template shortly after).
2. And Rich Farmbrough, isn't it a bit curious that you complain "could have been done" while you yourself added edits unasked for, while knowing that they rise discussion? -DePiep (talk) 09:26, 26 March 2012 (UTC)[reply]

Would you mind making substantive rather than ad hominem arguments? — kwami (talk) 09:52, 26 March 2012 (UTC)[reply]

You muddy your arguments. I suggest that before you post here, you edit out the junk and just post the sensible stuff. I know you are capable of advancing good points.

1. That is precisely my point. I can not only be more selective and keep happy those who think an unreferenced article is good, but one with an empty references section is bad, but actually add the reference as well.

2. No it's not curious. And the edits are not "unasked for". A job needs doing, it will get done, manually or mechanically. Many times jobs have been finished while people are still arguing the toss here.

Rich Farmbrough, 15:40, 26 March 2012 (UTC).[reply]

1. Now I get that the ref= input is to be entered manually by say Project members, because it needs a check with the main reference Ethnologue for live languages. Kwami expects 20% (~700) of the pages to fail (not to be referenced with E.), possibly because they are extinct or other exceptions (no ISO-3 code is another reason?). If I am correct with this, we could do this. Step1: When the infobox has no ref= input AND not extinct AND has ISO-3 code: THEN the bot adds ref=e16 and the Reference section (if not there already). Step2: Project members check each referenced language page, and delete or correct any wrong "ref=e16" (and with it the possibly emptied ref section). With this, on the minus side is that a wrong reference will be on pages for some time (estimate: less than 700 pages), and this check depends on future "promised" edits. -DePiep (talk) 18:20, 26 March 2012 (UTC)[reply]

Adding. To skip the drawbacks, we could do this: Step 2: the bot adds ref=eth16 by default as described, but not a ref section. Step 1: The template ref creation code is prepared to keep this one blank (no ref added), but adds the tracking Category:Language reference to be checked. Step 3: Project members check all category pages, and delete/correct the ref=eth16 when needed. When eth16 is correct, page stays unchanged. When all pages are checked, Step 4, by bot: with all pages that still have a "ref=" input, it makes sure a reference section exists. Step 5: the prepared /ref generation code is changed into "create reference for ref=eth16", which is the same as e16. The End: eth16 references are added and have a reference section available for sure. "eth16" stays as an extra option equal to e16. -DePiep (talk) 18:36, 26 March 2012 (UTC)[reply]

No, the bot should not add 'e16', because then we will be adding false refs. The refs need to be added manually. But the bot can add the infrastructure, so that it's as easy as possible for us to add the refs manually. Currently we need to add |date = 1987 |ref = e16 .... ==References== {{reflist}}. With this bot request, all we'd have to add is the 1987 and the e16. When you're going through hundreds of articles, that makes a huge difference is the time expended.

The point of this bot request is to make it as easy as possible to add refs, in order to encourage people to do so, and to make a project-wide effort as painless as possible. Many of these articles have been to all appearances unref'd for years now. — kwami (talk) 19:52, 26 March 2012 (UTC)[reply]

All fine with me, except that we do not want empty reference sections added. In the steps I described in the addition, the ref code is added by the bot (ref=eth16 in my example), but the ref creation is silenced temporally in the template (so no false refs are generated, and surely not the red warning). After a manual check of all pages, with removal of the ref code or adjusting wrong ones (that are not e16)). Only then a truly required ref section is added by the bot, and the template is changed to create the reference as expected. So the page check you describe is performed, but with a negative (deleting) edit. The words all pages might make this proces undesired though, since it is a sort of deadline requirement to finish the manual sweep some time. Of course, this is just one procedure we could do, one can prefer an other. -DePiep (talk) 20:31, 26 March 2012 (UTC)[reply]

I don't understand. Are you saying we should remove the refs currently in the articles? Also, the info box has no control over the ref section, so I don't see how your proposal could work. But this isn't the place for new proposals: better on the project page. — kwami (talk) 20:52, 26 March 2012 (UTC)[reply]

Retry: First we change {{Infobox language/ref}}: when ref=eth16 is entered, then no reference output is returned (blank), but a hidden tracking category is added. Then, all pages Language pages that currently do not use ref= in the template (article page), the bot adds ref=eth16 to the page (some ~3000 pages). Then Project Language members can manually check each page in that category: what ref to use? If eth16 (for e16) is OK, then no edit. Otherwise: change into appropriate reference like 'e15', or delete the ref input. When all pages are checked (weeks?), the bot does this: if a ref= is used on a page, then make sure there is a ===reference section===. After this, we change {{Infobox language/ref}} into normal: produce the reference, when input ref=eth16. When we save this edit, all template transclusions will show the reference correctly. eth16 is added for this process, and in the end does exactly the same as e16. -DePiep (talk) 22:53, 26 March 2012 (UTC)[reply]

What he's suggesting is this:

• Phase 1: add eth16 to all articles (auto) but suppress the display
• Phase 2: check and remove 20% that are invalid (manual) (this being less editing than adding the 80% manually
• Phase 3: add the ref section where needed (auto)

Somewhere the referencing would need to be turned on, though. Either before phase 3, which woudl give the big red errors, or after which would give transitional empty parameters.

It's basically a good idea as far as it goes but I'm not sure that the actual editing is the work, I think it's the checking. <meh> I also think that getting on with it would clarify the situation. Rich Farmbrough, 21:17, 26 March 2012 (UTC).[reply]

Correct. Of course turning the referencing on is done right after the bot has added reference sections where needed. This prevents red error texts ("ref section needed").

Note: if we want to prevent a permanent second code eth16 for its equal e16, we could turn off 'e16' referencing for the checking period, but that would leave current correct references blank. Better idea: the new code should use capital-E: ref=E16. First we check for the capital (checking period, will fill the category), but in the end we treat them E16=e16 giving the reference.

Note: adding and sorting parameters can be done as requested, in the first bot run.

Question to kwami: Switching on the reference production for the template (the final step), can only be done when all ~3000 pages in the tracking category are manually checked (because it will produce a reference on the article page. That ref may be false/unchecked = bad). You think that sweep can be done in overseeable time? -DePiep (talk) 22:53, 26 March 2012 (UTC)[reply]