Wiktionary:Grease pit/2008/October

Automate ise/ize variants?
I sometimes create missing ise/ize variant entries manually. In the absence of any wiki mechanism for indicating "this s can be a z" within one article, would anyone appreciate a bot to look at existing English ise/ize words, check for the alternative in some popular word list (or even Google, taking a large result set to mean it's probably a word), and create the "alternative spelling" article if appropriate and missing? Although I'm a fairly competent programmer and wouldn't mind attempting such a project, I haven't worked on a wiki bot before, so somebody of more experience might be able to put it together in minutes. Just an idea. Equinox 00:03, 7 October 2008 (UTC)


 * If a word is redlinked in an "Alternative spellings" section, then it might be possible to do something like Conrad.Irwin's accelerated creation of plurals and other form-of entries. I'm no expert though so I don't know if it actually would be doable Thryduulf 14:56, 14 November 2008 (UTC)

Template:en-adj-notcomp
It's time we orphaned and deleted this template. --EncycloPetey 18:13, 8 October 2008 (UTC)


 * Agreed. -Atelaes λάλει ἐμοί 19:05, 8 October 2008 (UTC)

Should terms used and linked in a defn be listed again in "Synonyms" or "Related terms" or "See also"?
I dislike redundancy and avoid listing a term as synonym or in other subsections if that same term has already been used and linked in the definition. An example I noticed today is epistemological. Defn1 for epistemological contains the term epistemology. Today an editor also added epistemology to the "Related terms" subsection. Another example is simian, where the definition includes the terms ape and monkey (both linked) and the "Synonyms" section also includes ape and monkey. Do we have a preferred practice here? Should terms like these be double-entered or is one usage in the definition all that we need? -- WikiPedant 19:52, 9 October 2008 (UTC)


 * For long entries (more than one screenful), I'd think that the duplication factor is not as important as the navigation factor. Duplicating near the bottom what appears at the top gives the user a chance to find something good to jump to on another page without having to scroll up or remember to scroll up. For such a short entry I wouldn't have put it below. I wouldn't get into a wheel war about it, one way or the other. DCDuring TALK 20:53, 9 October 2008 (UTC)


 * Since we're not paper, of course. If a related word is only in the definition, a future editor that rewords the definition, removing the related word, might forget to add it to one of the lower section. In addition, one might want to analyze our entries in a more structured way (eg pulling out synonyms and antonyms). This is only possible we list related terms in the lower sections regardless of their inclusion in definitions.  --Bequw → ¢ • τ 04:30, 10 October 2008 (UTC)

Should separate entries be created for 2-word (adj + noun) expressions where the adjectival entry already contains an applicable definition?
Today I was contemplating creating an entry for loaded question, but I noticed that loaded already has a sense tailor-written for this expression. I also notice that there is a tailor-written sense of loaded for loaded dice but that we do have a separate entry for loaded dice. Do we have a preferred practice in this sort of situation? On the one hand, I don't like to create a redundant--or sort-of redundant--entry, but, other other hand, a user might try to look up the expression "loaded question." What to do? -- WikiPedant 20:02, 9 October 2008 (UTC)


 * We could stand to have both, but the word always seems the more important to me for searching. If the sense of the word is really specific to that one idiom (if indeed it is an idiom), I'd favor a minimal sense and a "See" link to the idiom. If it is not limited to the idiom, but the idiom is very common, I'd favor a usage example including the idiom. Less common: 2 usage examples. If there is something special about the idiom, usage note, date, context, attestation, etc., then the user should be encouraged to go to the long idiom entry instead of settling for something short at one of the component words. Just an opinion, though. DCDuring TALK 21:02, 9 October 2008 (UTC)

Misspelling template
So, I've had a few problems getting the code to work with it. Upon bracketing the target word, comes out as "Common misspelling of  [[raspberry|raspberry ]] ." Could someone tech-savvy fix Template:misspelling of? Teh Rote 22:02, 10 October 2008 (UTC)
 * This is deliberate- we don't want wikilinks on a misspelling page because we don't want it to count in the total number of pages. Nadando 06:02, 11 October 2008 (UTC)
 * Huh? The template hard-codes a link... I'm not sure what you're talking about. Mike Dillon 06:34, 11 October 2008 (UTC)
 * The entry-county software counts any entry with double brackets as real. Misspellings are supposed to be outside the count. The appearance that you obtain is intended to discourage cheating the counting system. After all, one could claim that huba:, buba:, etc were all misspellings of tuba: and thereby inflate the entry count, but not the value of Wiktionary. DCDuring TALK 12:53, 11 October 2008 (UTC)
 * Are you saying that only pages with literal wikilinks in them (not inside a template) are counted toward the entry count? That's pretty cheesy... Mike Dillon 19:14, 11 October 2008 (UTC)
 * It is even cheesier than that: it is entries with "Robert Ullmann 12:14, 13 October 2008 (UTC)
 * Maybe I am missing the point but ... why would you want to write when you can achieve the desired outcome (an internal link to raspberry:) by simply writing  ? It can't have anything to do with counting, right, Robert? -- Gauss 23:41, 12 October 2008 (UTC)
 * I think Teh Rote is just used to linking the parameter for form of templates, and is surprised that this one (intentionally) doesn't allow that. The answer is, just don't link it.


 * Would be much better if the s/w would just count main namespace entries that did not contain __NOCOUNT__ or some such. There are multiple bugs entered for this, no sign of anything happening. Robert Ullmann 12:14, 13 October 2008 (UTC)

It gets worse: the current count is updated (incremented/decremented) based on the presence of brackets vs the previous version of the page. But the updateArticleCount script uses "actual" outgoing links. So when they ran the script recently after server problems caused the counter to stop for a while, it overcounted by some amount.

In fact, there are 986,887 entries in NS:0 as of a few minutes ago; the counter reads 972,692. Robert Ullmann 12:48, 13 October 2008 (UTC)
 * If this is correct, am I right in saying we are now over 1 million? (14 k plus the current count). Nadando 05:04, 16 October 2008 (UTC)


 * As of yesterday's XML dump (15 Oct) we were just short. Today may be over 1M. Robert Ullmann 11:55, 16 October 2008 (UTC)

template:homophones
Could someone with the necessary skills adjust the template to add optional qualification parameters in the template. Currently it just takes up to 5 positional parameters that link to other entries. This works fine where up to 1 needs qualification, as this is just placed last and the template is used. However it doesn't work where more than one link needs qualification, e.g. at wicker the following is needed: *   *   giving the less than ideal

What would be better would be something like: * to give:

Actually, as these are the most used qualifications, it might be worth standardising them somehow, maybe something like: * to produce the same as above. Obviously Wicca and whicker are homophones of each other non-rhotic accents with the wine-whine merger, so Wicca might get * to give Thryduulf 10:50, 11 October 2008 (UTC)


 * As there is no way of telling in which order named parameters are given, this could be very tricky. It might be better to have a seperate template (maybe ) that can be appended after :

*   * Homophones: Using the r= parameter to give a short-hand reason. Conrad.Irwin 13:41, 11 October 2008 (UTC)

Broken categorization?
I have added Categories to the pages collar, marisco, mariscos, mariscal about 30 minutes ago, however the pages still not appear on the Category pages. Might it be that categorization is broken somehow or is the server just to slow? Matthias Buchmeier 19:50, 12 October 2008 (UTC)
 * The entries appear now in their categories, at least for me. So it was probably a consequence of the major server problems we had today. -- Gauss 23:54, 12 October 2008 (UTC)

New messages template won't go away
Why is it that the new messages template won't go away even when I check my IP talk page? Could the fact that the school district uses three different proxies on completely different IP ranges that alternate at random have anything to do with it? 169.139.98.194 11:52, 13 October 2008 (UTC)


 * You shouldn't get any given one more than once; but you might see several different ones. Simplest thing would be to create an account? Robert Ullmann 12:05, 13 October 2008 (UTC)


 * I once made the mistake of editing wikipedia anonymously from a friend's AOL account. I think I went through about 20 message pages before I worked out what was happening. Thryduulf 16:57, 14 October 2008 (UTC)

Han character formatting
Just a note re what I am doing; I, or UllmannBot, have been lax in keeping up with the set of entries for Han characters. I hadn't run the problems report since January, and it went from ~900 to ~2100. A large number of these are from an IP-anon (the one who refuses to use his login) adding Vietnamese without using the template (which he knows perfectly well). Another large set are Korean added w/o the "stub" definition line.

I am running some automation to go through User:Robert Ullmann/Han/Problems and fix the things it has been taught to. It is also adding sort keys to calls, which makes those cats a bit easier to deal with (else each character is listed under itself, not terribly useful ;-). Robert Ullmann 14:18, 17 October 2008 (UTC)

redlink categories?
Is there a report that lists non-empty categories that don't have a category page? RJFJR 14:44, 22 October 2008 (UTC)


 * Special:WantedCategories Robert Ullmann 14:47, 22 October 2008 (UTC)

Some automation broken at 04:30 UTC this morning
If something of yours stopped working, this is the explanation:

They made a change to the MW software to require the wpSection field be supplied on an HTTP POST to edit a page, even if nil (as it would always be if not section editing. This was a completely gratuitous change.

Worse: it does not return an error when the field is omitted; it just returns the preview page, as if action was preview. So it is very non-obvious what has happened.

So if post (edit) operations that worked before today are now failing, this is likely the problem. If you need help fixing this, please ask me. If you are using a sync'd version of the python wikipedia framework, re-sync it; they have patched the problem.

Sigh. Robert Ullmann 14:32, 25 October 2008 (UTC)
 * Hi Robert. SemperBlottoBot has stopped working - How do I "re-sync" it? (or otherwise fix it) SemperBlotto 16:26, 25 October 2008 (UTC)


 * If you used svn checkout to get it in the first place, do it again. (After making sure it won't overwrite any modified files in place; e.g. if you modded pagefromfile.py without changing the name.)
 * If you downloaded a zip file or something, do that again.
 * Or: edit  and look for lines like this:

predata = [ ('wpSave', '1'), ('wpSummary', comment), ('wpTextbox1', text)]


 * this will be the first occurrence in the file of, in a routine called   in older versions, and   in newer versions. Change it to add  :

predata = [ ('wpSave', '1'), ('wpSummary', comment), ('wpTextbox1', text), ('wpSection', '')]


 * this is in older versions, in newer versions, it is a dict, and it should look like this:

predata = { 'wpSave': '1', 'wpSummary': self._encodeArg(comment, 'edit summary'), 'wpTextbox1': self._encodeArg(text, 'wikitext'), 'wpSection': '' }


 * this should give you some idea of just how crappy the change they made was: it is failing because you aren't supplying an un-needed parameter with a null value. Pfui. Ought to simply be fixed on the WM side. (If you can just edit in the patch, I'd recommend that way of fixing it, as everything else you have is working; no point breaking it. ;-) Robert Ullmann 17:12, 25 October 2008 (UTC)


 * I have editied wikipedia.py as above (first version) and saved it. When I reran the bot it rebuilt wikipedia.pyc (with no errors) but the bot gets the same errors when it tries to add a page. (Just to prove that the edit was good, here is the part of wikipedia.py copy/pasted)

text = text.encode(self.site.encoding) predata = [ ('wpSave', '1'), ('wpSummary', comment), ('wpTextbox1', text), ('wpSection', '')] # Except if the page is new, we need to supply the time of the # previous version to the wiki to prevent edit collisions if newPage: Any ideas? SemperBlotto 18:01, 25 October 2008 (UTC)


 * Hmmm, that's the only change I made. The only difference I can think of is that I'm editing pages and you are creating them. Let me check this out a bit. Robert Ullmann 18:22, 25 October 2008 (UTC)
 * Good news and bad news: the good news is that that is the difference; creating pages still fails for me; the bad news is that I don't know why yet ... Robert Ullmann 18:36, 25 October 2008 (UTC)
 * See bugzilla:1181 and r42037, they added a gratuitous check, just as I thought. It breaks page creation for everyone, even with the current pybot. The problem had already been solved by re-ordering the form. Should have the other part of the patch for you in a few minutes. (testing) Robert Ullmann 19:38, 25 October 2008 (UTC)
 * Okay: a few lines below that, change to:

if newPage: predata.append(('wpEdittime', '1')) predata.append(('wpStarttime', '1'))


 * that is, change the null value to "1". This will work if the page never existed; if it has, and has been deleted, you will get an edit conflict. We'll see what real fix "they" come up with. Robert Ullmann 19:53, 25 October 2008 (UTC)


 * Better: eliminate the test for, always set the parameters as in the else case:

predata.append(('wpEdittime', self._editTime)) predata.append(('wpStarttime', self._startTime))


 * that was the change added to the framework a few hours ago.


 * meanwhile Brion is raging once again about people using the GUI instead of the API. Never mind that the edit API has only been available for a few weeks. (*sigh*) So there are many thousands of people out there with this code.


 * he is taking the superfluous checks out of the MW s/w, but that will take a while to commit and install. Robert Ullmann 20:27, 25 October 2008 (UTC)


 * Thanks - I went with your latest suggestion (without the check) - everything now works again. SemperBlotto 21:31, 25 October 2008 (UTC)

Editing from two different userids at the same time.
For some time I had been convinced that two of our regular users were manifestations of our old friend Wonderfool. Then I noticed both of them editing at the same time, including two new terms added within seconds. So I tried an experiment. I had two sessions open in Internet Explorer, and logged off one of them and then logged on as a second user (SB2). At my next usage of the other session it had changed from SemperBlotto to SB2 - so use of two userids at the same time was obviously impossible. I then fired up Google Chrome (Firefox would probably have worked as well) and logged on as SemperBlotto. I was then able to edit as SB2 on Internet Explorer and as SemperBlotto on Chrome at the same time, and added two terms within seconds.

I can't think of any way of preventing this, as the two pieces of software seem to hold their own versions of the same cookies. SemperBlotto 10:08, 27 October 2008 (UTC)


 * I could run FF, IE, and Chrome with 3 different logins. One can also have multiple copies of FF or Chrome, or have multiple log in sessions on the same machine, each with its own cookies. (And I edit as me, UllmannBot, AutoFormat, and Interwicket at the same time ;-). Robert Ullmann 14:28, 27 October 2008 (UTC)

Template:cs-verb
In Template:cs-verb, I want to add a parameter so that if that if all displayed is just, then it would add to a category Category:Czech verbs not classified as perfective or imperfective. I managed to add a parameter to divide them in 2 categories for perfective and imperfective, but I'm having trouble with this newer one. --Ro-manB 15:21, 27 October 2008 (UTC)


 * It can be done with nested #ifs as you were trying to, but easier to use #switch again, and just put that cat in the default case. I did that, and also wrapped it in a check for namespace 0; take a look, I'm sure you'll understand what I did. Cheers, Robert Ullmann 15:33, 27 October 2008 (UTC)

System crashes trying to go to WT:LOP
Trying to get to WT:List of protologisms (or the link to WT:LOP on the edit screen that opens for the red link) pretty reliably crashes Fedora core 6 running Firefox 3.0.3 here, leading to the Fedora user login screen. After restart, Firefox successfully reinstates the session when asked to do so. So far that's the only page I've tried to get to that does that. __Just plain Bill 18:03, 29 October 2008 (UTC)


 * In the absence of a set of qualifications for addition to that list, the thing just keeps growing. I move we delete the whole thing.  It adds nothing to a respectable, usable dictionary.  Perhaps UD would like a transwiki of it.  -Atelaes λάλει ἐμοί 19:31, 29 October 2008 (UTC)


 * Is UD under the GFDL? Thryduulf 21:56, 29 October 2008 (UTC)


 * I do believe everything on Wiktionary is GFDL. -Atelaes λάλει ἐμοί 05:58, 30 October 2008 (UTC)


 * I think you misread Thryduulf's question? As you say, everything on Wiktionary, including LOP, is GFDL; so UD can only take a transwiki of LOP if UD is GFDL-compliant. —Ruakh TALK 18:59, 30 October 2008 (UTC)


 * Clearly I should have spent an additional microsecond reading that comment. Sorry.  UD does not appear to be GFDL, so no, I don't think they could take a transwiki (I was being largely sarcastic in that comment anyway).  -Atelaes λάλει ἐμοί 03:54, 31 October 2008 (UTC)


 * I think a lot of users do make good use of LOP; the problem is with a few users who go crazy, e.g. adding 50 million "go the way of " expressions or 50 million "red/blue " expressions (with red=Republican, blue=Democrat) or 50 million adjectives in "-orse" and "-orst" (opposite of "-er" and "-est"). It would sadden me a bit if we deleted a widely-used project page just because of a few users who don't play well with others. (But you're probably right that the page doesn't really add anything, and I'm not prepared to fight for it.) —Ruakh TALK 18:59, 30 October 2008 (UTC)


 * I'm curious as to what "good use" would be. If, by good use you mean adding words in reasonable numbers that they've spent a little time thinking about, then I would be prepared to concede.  However, I would respond by asking what good such good use does.  I think that the vast majority of words on that list are words with virtually no usage.  Now, they might be creative and interesting (and quite frankly, most of them aren't), but they're not descriptive of the language.  If WT:LOP had some sort of criteria, say a third of CFI, that might be interesting and useful, as it would be a list of words that have a decent shot at getting into the main namespace in a few years, and thus words to keep an eye on.  It would be an efficient way of keeping Wiktionary on the cutting edge of language, something I support.  However, as it stands, I think it don't do no one no good.  -Atelaes λάλει ἐμοί 03:54, 31 October 2008 (UTC)

This issue may not be specific to WT:LOP since now trying to go to WT:GREASE from the main page crashes my Fedora session. __Just plain Bill 02:16, 30 October 2008 (UTC) and, it may be a corrupt user account at this end; only seems to happen with the one. __Just plain Bill 03:43, 30 October 2008 (UTC)


 * As Atelaes alluded to, I think it has to do with how huge the pages in question are. —Ruakh TALK 18:59, 30 October 2008 (UTC)

I tried to make use term:, which worked reasonably well (I only needed 3 edits yet :-) ), but there is one thing that I do not manage: to make the second transliteration show up.  Can someone have a look at this? I think simply removing the ss parameter will be the easiest to do and I don’t think it is that useful anyway.
 * I there a way to specify template arguments in Special:ExpandTemplates? That would make it unnecessary to save changes just to look what the result looks like. H. (talk) 14:08, 30 October 2008 (UTC)


 * You shouldn't be experimenting on the live template at all. Create (or ... 2 or whatever, and try to make it work. In this case, it should be restored to the working version. Then consider whether using term is a good idea: IMHO, almost certainly not. Is already overloaded. Robert Ullmann 18:01, 30 October 2008 (UTC)


 * While I acknowledge it is better to create a temporary template for experimenting (hadn’t thought of that, sorry), I do not agree that it should be restored. I think it is an elegant solution to delegate to a template that was made for this. H. (talk) 09:20, 31 October 2008 (UTC)


 * As a result of this change, there are loads of new suffixy categories which need to be added - see WantedCategories. How should these new categories be categorised? --Borganised 11:22, 6 November 2008 (UTC)


 * Actually, these changes are independent, although I made them both. Why don’t you go ahead and make some template (not necessarily a wiki template, just a sketch, I mean) for one of those categories, which can then be copied to the others?  Something like “This category contains words that end with the suffix ‘-X’.  See, Category:English suffixes.” but with some more layout. H. (talk) 11:43, 6 November 2008 (UTC)


 * Would it be too late to as that the quote marks be removed? They are untypable, ugly and unnecessary. Conrad.Irwin 01:44, 12 November 2008 (UTC)


 * What do you mean? Which quote marks?  They are not untypable with en-us intl on Linux, and even if you cannot use that, you can use Edittools, under Misc.  Why are they ugly, where are they unnecessary? H. (talk) 17:42, 18 November 2008 (UTC)