User talk:DerbethBot

From Wiktionary, the free dictionary
Jump to navigation Jump to search
This is a talk page of a bot (a mindless program), so there's very little sense in writing anything here. Don't expect a reply if you post here. Please use my talk page instead.
--Derbeth talk 14:53, 12 January 2015 (UTC)[reply]


Cool. Does the bot understand that if there is more than one ety section, the pronunciation should usually be before them, at L3? Robert Ullmann 23:25, 26 November 2007 (UTC)[reply]

When there is more than one Ety section, and the pronunciation is the same, WT:ELE doesn't cover the case; we put the pronunciations first. Putting it after/inside the first section is wrong. And will require cleanup. Has it already done this? Do you know? Robert Ullmann 11:15, 27 November 2007 (UTC)[reply]
If there is no Etymology section, you put Pron first, or after Alt Spell; if there is an "Etymology" section put Pron after it; if there is an "Etymology 1" section" you put Pron before it. Are we going to have to run checks for Ety 1 in every entry you touch? ScsRhymeBot made a mess of this. You are less likely to see multiple etys, but please look for "Etymology 1". Robert Ullmann 11:20, 27 November 2007 (UTC)[reply]
It is not "insane". Are you checking for "Etymology 1" or not? Robert Ullmann 22:17, 27 November 2007 (UTC)[reply]

block[edit]

Derbeth: this is not acceptable.

You must make your code recognize that a new pron sect goes before Etymology 1 (or not added at all)

You must TEST it and see that it works. "hope" is not sufficient.

Saying the requirement that your bot work properly is "insane" is not encouraging, it is direct evidence that you are too careless to be operating a a bot.

Note: what the bot is doing is good (great!) but if you are unwilling to TEST it and make sure it WORKS CORRECTLY, you must stop.

Also: I'm not (was not) asking you to prove it works correctly, only that you tell us it does. But at this point, even that level of trust is very nearly erased.

Show that you understand the issue, that you take it seriously, that you have fixed the code, and we can continue. Unless and until, this bot must remain blocked. I'm sorry, this is an excellent thing. It is a crying shame that you dismiss problems. Robert Ullmann 23:02, 27 November 2007 (UTC)[reply]

Note to everyone else: the issue is the placement of a new Pronunciation section in the presence of multiple Etymology sections within the language section. Either it should be before all; or simply not done at all. ScsRhymeBot ignored this issue and there are still many entries that need to be manually corrected. Derbeth is correct that this is "unlikely" for languages other than English, but it does occur, and at the time of the block, the bot was doing English entries. Derbeth has offered no statement (let alone evidence, which was not requested) that the bot code checks for this condition, either adds in front or skips or tags, except "hope" that it works, which clearly indicates to me that the code has not been tested.
The reference (on my talk) to this issue being "insane" is troubling. Robert Ullmann 23:11, 27 November 2007 (UTC)[reply]
This edit and this edit made just before the block are both incorrect, and indicate serious problems with the bot code's logic. (note the addition of a new Pronunciation section in both cases, where the existing section was placed correctly, although the Alt Spell section was not; the placement of the new section is in any case incorrect) It needs to be recoded/designed/whatever and tested on these cases. Robert Ullmann 23:38, 27 November 2007 (UTC)[reply]
Any admin feel free to review/modify the block. My sleep time. Derbeth, I am very sorry about this, I hope we can sort it out. Robert Ullmann 23:38, 27 November 2007 (UTC)[reply]

this edit is also a problem, although the entry was not structured correctly. Robert Ullmann 12:07, 28 November 2007 (UTC)[reply]

Worked correctly at bit for cs and sv. (Though still needing work because the audio includes the article ;-). Robert Ullmann 12:11, 28 November 2007 (UTC)[reply]

should work like this[edit]

One reasonable sequence:

  1. extract proper language section; then within that:
    (if no audio there already:)
  2. if multiple Pron sections, STOP
  3. if one Pron section and 0 or one Ety section, add audio
  4. if one Pron section and 2+ Ety sections and Pron before the first, add audio
  5. else if Pron after the first Ety, STOP
    (so no Pron section, continue:)
  6. if 0 Ety sections, add Pron after Alternative spellings (or forms) (careful of levels, before next L3), or at top
  7. if one Ety section, add Pron after it (again, before next L3)
  8. if 2+ Ety sections, add Pron just before first

The alternative to "STOP" is to add the Pron section with a cleanup/RFC tag so you can go and place it correctly; or someone else will eventually get to it. A somewhat sloppier alternative is to always place a check tag if there is an Etymology 1 header. (But note the bot code misbehaved at theater/theatre, where there was only one.)

The existence of Etymology 1 is a reasonable heuristic for "multiple Ety sections", but since you must count Pron sections, you might as well count both. Start by setting ety=, pron=, then you can proceed with code like: (step 4 above)

if pron == 1 and ety > 1 and text.find('==Pron') < text.find('==Ety'): [add audio to Pron sect]

and so on. This is all very good to do, when done right. Robert Ullmann 10:55, 28 November 2007 (UTC)[reply]

block issue[edit]

(from my talk page)

I have always tested my bot after every change in the code, but no testing can prove that the bot will always work fine in all cases, as you can't predict any abnormal entry form. When I have written "I hope" I have ment "I have tested my bot and therefore I can hope". Well, but know better than me what I do.

"test" doesn't mean try one and hope it continues to work. If you have tested something, you don't "hope", you "know" (with some degree of certainty not quite 1.0) that it works. Robert Ullmann 10:27, 16 December 2007 (UTC)[reply]

What I find "insane" is that noone on English Wiktionary has been able to tell me how exactly my bot should work, although it was about a month from my first post on Beer parlour to the moment when my bot was given a bot flag. I have clearly written how my bot works and noone has objected; noone has ever mentioned sections like "Etymology 1". And in the middle of the work of my bot you have written on its discussion page, that I should completely change its behaviour, because it was completely wrong. When I said that I had written the bot according to Wiktionary official rules, you replied that these rules are wrong. How can I write a bot when noone on the English Wiktionary actually knows, what is the structure of entries?

I mentioned multiple etymologies immediately on your Beer Parlour posting; you replied dismissing it as an issue; I replied again saying it was an issue that MUST be addressed. Which you did not respond to. So your statement here that "noone has objected" is demonstrably false.
I did not say that ELE was wrong; I said it was incomplete; the case of multiple etymologies with one pronunciation is not covered. We do know what the structure of entries is; your not listening is an issue. And, as noted, this was raised at the very start. Robert Ullmann 10:27, 16 December 2007 (UTC)[reply]

Even the algorithm you have provided me is not right. The edit on theater you have found wrong is actually the only possibility: pronunciation section is in wrong place, because the rest of the entry is being transcluded as a template, so my bot cannot reach it. Even if I dealt with the case of (multiple?) page transclusion (which would require major code rewrite; I don't have enough time to do such complicated job), I think that soon an another bizzare case similar to "Etymology 1" or "page transclusion" would appear. I think that (lack of) organisation of English Wiktionary makes any automatic processing impossible as there are more exceptions than rules.

I have no trouble at all with automatic processing; it is not intractable; it just has one level more structure than the other wikts typically do. (you are correct about theatre/theatre, that is/was an example gone wrong.) The algorithm is fine; if you actually follow it and the entry is already malformed, that is not your problem. Refusing to follow it is a problem. Robert Ullmann 10:27, 16 December 2007 (UTC)[reply]

Currently I don't have time for anything; I am not able to implement complicated-as-hell handling of every possible exception on English Wiktionary. I see two solutions: either I will forget about audio files on English Wiktionary or I will modify the bot code, so that it would add pronunciation in a simple but safe way (in order not to remove anything) and would insert a cleanup template, so that people can fix them later. I think that for most languages except English such job would be just removing the template. There are more than 10 000 audio files to add, the choice is up to you. --Derbeth talk 19:14, 9 December 2007 (UTC)[reply]

No, adding cleanup tags because you are unwilling to make the bot function properly is not acceptable. (Adding a tag when the bot can not reasonably know what to do is fine, and a good idea.)
It you make the bot function properly, and properly test it, fine.
If, OTOH, you maintain the attitude that you can ignore or dismiss the format and policies here as "insane" or too complicated, then you should, as you say, find something else to do. Is a crying shame, as it would be very useful. Robert Ullmann 10:27, 16 December 2007 (UTC)[reply]
Other sysops: I've left the block in place; feel free to lift it if you can get any kind of assurance from Derbeth that the bot will function properly. I'm not hopeful in that regard. Robert Ullmann 10:27, 16 December 2007 (UTC)[reply]

bot malfunctioning[edit]

Is adding plural pronunciations to singular enties. Blocked for now. Please fix it, and take more care. Robert Ullmann 15:43, 20 November 2008 (UTC)[reply]


It is also added links to bad files, such as the "Canadian" pronunciation of about, which was deliberately not linked as a fatuous pronunciation. That entry alreay has a correct Canadian audio link. --EncycloPetey 23:04, 20 November 2008 (UTC)[reply]

Homonyms[edit]

The following edits need to be checked - they're on a page of homonyms.

The bot added a pronunciation to this page, but it's a misspelling, not a word in English. --EncycloPetey 21:11, 12 September 2009 (UTC)[reply]

Commons doesn't work that way. I can't request an "incorrect" pronunciation be deleted there. They don't work that way. I did request a page move for their ad nauseum file to the correct spelling, but any such change could take weeks to happen. This is still only one instance, and meanwhile the problem will have to be corrected via your bot somehow. You will have to take responsibility for the erroneous edits of your bot, as it is a requirement of our bot policy. If you are unwilling to abide by Wiktionary policy, the permissions granted to run a bot here will have to be reconsidered by the community. --EncycloPetey 15:43, 16 October 2009 (UTC)[reply]

Help with the addition of a file pronunciation, in other language sections[edit]

Hello, could you share the source code of your bot User: DerbethBot.

Specifically, the functions for adding the pronunciation based on the availability of pronunciation of Esperanto in this category http://commons.wikimedia.org/wiki/Category:Esperanto_pronunciation.

We very much need to add pronunciation articles Esperanto, in Russian Wiktionary (and later in vikivortaro).

I regularly add audio files of pronunciation of Esperanto in this category http://commons.wikimedia.org/wiki/Category:Esperanto_pronunciation

But I do not have time or do I need a lot of extra time trying to prescribe is the pronunciation of all articles, you could help me?

That's how I add the pronunciation articles http://ru.wiktionary.org/w/index.php?title=azeno&action=historysubmit&diff=1507519&oldid=1496372 manually. In the Russian Wiktionary.

And yet, I apologize for my English. Some suggestions I have translated by machine, unfortunately I do not know Polish, but you do not know Russian or Esperanto, because will be in English explained. Balamutick 13:25, 12 October 2009 (UTC)[reply]

The French Wiktionary needs you[edit]

I apologize to ask you to do the same job of here on the French Wiktionary, which is now waiting for it in its grease pit for 3 days: actually the wonderful Shtooka association which had recorded 100 000 free sounds, including some that you've already deployed, has asked for a presentation of the French Wiktionary in Paris the next 13th February, and it will be more constructive to do it with their sounds inside. JackPotte 00:31, 30 January 2010 (UTC)[reply]

Simple English Wiktionary[edit]

Hey, when your bot adds audio files, could it place the UK file in front of/above the US file so that we can keep it in alphabetical order? Thanks for your attention. Tygrrr 19:10, 16 April 2010 (UTC)[reply]

It's great that your bot is adding Mandarin pronunciations, but it would be even better if it could add the pronunciation for both the traditional and simplified form (as in 高級 / 高级). ---> Tooironic 11:55, 7 December 2011 (UTC)[reply]

Another example is 說服 and 说服; you didn't add the pronunciation for the former. ---> Tooironic 21:14, 15 December 2011 (UTC)[reply]
And, again, at 证实 and 證實. Come on! ---> Tooironic 03:11, 19 December 2011 (UTC)[reply]
Plus at 慶祝 and 庆祝... ---> Tooironic 03:17, 19 December 2011 (UTC)[reply]
And at 臨時 and 临时. ---> Tooironic 22:17, 22 December 2011 (UTC)[reply]
And 正義 and 正义, and 嚴厲 and 严厉. ---> Tooironic 01:16, 26 December 2011 (UTC)[reply]

Latin pronunciations[edit]

On 18 Oct 2011, your bot added numerous "Roman" pronunciations to Latin articles. These need to be removed or relabeled, as these itmes are not "Roman" Latin. The filenames were erroneous, and most have since been corrected or deleted by their creator. All reference to the "Roman" dialect (as he called it) have been eliminated from Wikibooks, where it was originally put forward. This is a result of his discovery that his descriptions and explanations were in error. So, these audio files should not be linked here, and many of them are awaiting deletion on Commons. --EncycloPetey (talk) 23:24, 12 August 2012 (UTC)[reply]

Does this bot have an exception or exclusion list? The file Pt pagar.ogg is mislabeled- it sounds more like Italian pagare. Every time DerbethBot adds it to pagare, the edit has to be reverted. Chuck Entz (talk) 00:54, 10 December 2012 (UTC)[reply]

Your bot is making some mistakes[edit]

At خبز, the audio your bot just added is one that I had only just removed because it is mislabeled as the wrong word.

At أهلا, the audio your bot just added is the same one that is already there.

--WikiTiki89 16:57, 2 March 2014 (UTC)[reply]