User talk:MerlIwBot

From Wikispecies
Jump to navigation Jump to search

Removal of content[edit]

Your bot has been removing vernacular names from many pages. those edits need to be undone and corrected. We have users who sometimes miss-add the vernacular names in the form of an interwiki link because they do not understand the syntax. Your bot is interpreting these errors as bad links, and removing them instead of fixing them. See for example, this edit. --EncycloPetey (talk) 05:53, 14 July 2011 (UTC)[reply]

See discussion at Village Pump for further details.Koumz (talk) 00:35, 15 July 2011 (UTC)[reply]

More incorrect bot additions Adoxa/Anthocerotopsida[edit]

The latest changes made to Adoxa and to Anthocerotopsida are incorrect. In the Adoxa edit, links to a one particular species were added, instead of the pages for the genus (see the Polish article, for example). For class Anthocerotopsida, the articles on the division Anthocerotophyta were added instead of links for the class. --EncycloPetey (talk) 00:50, 22 August 2011 (UTC)[reply]

btw.: Because species isn't already scanned by templatetiger i still cannot create that missing VN entry list as discussed last time. Thats why i added a new feature to my bot last last week. Interwikis can be added or modified, but interwikis that are only removed must either have an existing vn template parameter in that language or the same page title. The other removals send to a queue for later.

Adoxa[edit]

The dewiki interwiki to de:Moschuskraut at Adoxa is wrong. It belongs to Adoxa moschatellina. So i would split this into these groups. I have moved articles about both (genus and species) to the species to solve the conflict.
Group 1
Group 2

Anthocerotopsida[edit]

A bit more complicated. My suggestion would be:
Group 1
Group 2
Group 3
Group 4

Merlissimo (talk) 07:48, 22 August 2011 (UTC)[reply]

Thanks for the update on the VNs. I think that strategy will work for that issue. (It has worked so far.) Koumz (talk) 12:53, 22 August 2011 (UTC)[reply]
Because there was no comment against my suggestion for a week i have changed the interwikis as suggested above. Merlissimo (talk) 15:38, 3 September 2011 (UTC)[reply]

I couldn't understand what the "groupings" meant, and so could make no comment. The most recent batch of changes are still very wrong, and have no been propogated to multiple wikipedias. I have reverted on the Indonesian Wikipedia, but don't have the time to fix the problmes globally. My experience is that, if I fix them, a bot will just come along and undo all my work, so the problem will only be solved if a bot is correcting the links properly. This needs to be sorted out across all those articles. --EncycloPetey (talk) 04:34, 4 September 2011 (UTC)[reply]

What do you mean by "most recent changes". Another automatic change or only the four changes to the articles above which was my human decision?
In an interwiki group all pages are linked to each other. Each interwiki group contains articles about the same topic. I have added id:Lumut tanduk to group 3 because i think it the same topic as e.g. species:Anthocerotopsida. Thats why my bot added it there [1] and removed it from species:Anthocerotophyta[2]. On idwiki all interwikis to another group were removed [3]. If you think the idwiki article belongs to group 1 and not to group 3 just tell me.
Interwiki relationship is a 1:1 relation by design definition. So you must not have two article at one wiki that link to the same article of other wiki. One day interwiki will be removed from all articles and stored in a central database. Until this extension is enabled on wikimedia you can tell me which articles belong to one valid interwiki group on my talk page here and i can propagate this to all wikis. Merlissimo (talk) 07:41, 4 September 2011 (UTC)[reply]
If interwiki relationship must be 1:1, then it cannot cross-link Wikispecies and the Wikipedia articles. The topical coverage is not 1:1 between pages, and while all linked pages should cover the same topic, sometimes one or more Wikipedias will have a combined article. So (for example), one article on the Korean Wikipedia may cover the same information as two articles on the English Wikipedia, and this may be covered by 3 articles on Wikispecies. The links are not 1:1, and that assumption can never work.
I still don't understand what the groupings mean, or why there were two "Group 3" collections. The Indonesian page id:Lumut tanduk is the same as Anthocerotophyta, not Anthocerotopsida. What is "Group 3" supposed to include? It makes no sense. And why are some items listed in more than one group? --EncycloPetey (talk) 17:52, 5 September 2011 (UTC)[reply]
The cardinality of the interwiki relationship is 1:1. This is part of the interwiki use cases that was discussed in 2004. So your linking is not allowed by this use case and not supported by mediawiki. You can create a feature request i you like to change this, but this would need many changes to mediawiki source code. Of course you can add those interwikis to page wiki code because nobody wrote an error message yet, but the back end database is not supporting this behavior (e.g look at the keys of interwiki table). And my interwiki bot must follow these use cases to be compatible to current mediawiki version.
Some people are working on strategy:Proposal:A central wiki for interlanguage links as an extension which would make interwiki bots unnecessary.
Except commons category no article is listed on more than one groups. commons category should link to article namespace and if there is no such article it links to category namespace. Thats why commons category is exactly part of one article group and one category group. Merlissimo (talk) 16:04, 6 September 2011 (UTC)[reply]

Clearly, the problems have not been sorted. [4]. Your bot continues to add links that are incorrect, even after repeatedly being informed of the problem. If it takes this much work, over many months, to sort out a single taxon's links, then I am convinced that a bot should not be used to link taxa across projects. They just can't do it right. --EncycloPetey (talk) 20:15, 11 November 2011 (UTC)[reply]

You can find my solution above. Because no one ever responded to my suggestion till today i have not fixed this problem manually. Please leave a comment if the above grouping is ok for you.
It also won't help if you readd interwikis to not existing pages. Merlissimo (talk) 02:55, 13 November 2011 (UTC)[reply]
I tried to discuss this above and pointed out that three groupings will not work. I asked questions about tis problem, and pointed out an underlying flaw in the the logic used for linking. Your response has shown that (1) the bot is not capable of dealing with the complex pattern of linking required between Wikispecies and Wikipedia, since your bot is limited to 1:1 matches, which will not happen. You can see, even in the meta-discussion you pointed out, other people working on the linking project understand that linking between wikis is not 1:1, but you persist in trying to force it to work that way. (2) You have little or no understanding of biological classification, and so cannot correct problems yourself. I'd say then that this bot experiment has failed and should be ended. --EncycloPetey (talk) 15:30, 13 November 2011 (UTC)[reply]
This is not because of my bot framework. It is limited by design concept of mediawiki. Currently some developers are hired, so that adding those conlict manually - as you did - won't be possible anymore.
I cannot understand, why you readded languagelinks to not existing page (e.g: he:אנתוצרוטיים), or why nn:Nålkapselmosar should be wrong although no:Nålkapselmoser has exacly the same content. Merlissimo (talk) 17:06, 13 November 2011 (UTC)[reply]
Why haven't you made any of those corrections? I simply reverted a bad bot edit; it is much easier to revert a complicated bad edit than to go through manually, item by item, to make specific corrections when so much incorrect editing has occurred. This is a wiki. If you spot a problem, then you can correct it.
The reason you don't underdstand why the nn is wrong is (as I said above) you do not understand biological classification. The nn page is about the entire Division Anthocerotophyta, not about the class Anthocerotopsida. The no link is incorrectly placed. This change obviously requires that the person making the change understand biological clssification. Your bot does not understand, and you do not either. Thus, your bot is making poor edits, and you are unable to make corrections.
What I don't understand is why each time you run the bot, it makes the same linking errors, even though the problem has been pointed out several times. This shows an unwillingness to solve known problems. I also don't understand why the bot isn't fixing article links. I just added he:אנתוצרוטה to Anthocerotophyta, which is a link that for some reason your bot never added, even though it has apparently been linked from the English Wikipedia for quite some time.
I also do not understand why you persist in believing that the edit structure of your bot is correct. It is not. The discussion on metawiki points this out, yet you continue to believe that 1:1 linking is possible and desirable. It isn't.
My conclusion is that your bot does not (and cannot) do what it was proposed to do. --EncycloPetey (talk) 18:05, 13 November 2011 (UTC)[reply]
It tried to solve this bug conflict many times. Reverting everything is not helpful because it doen't make to conflict smaller. I still think that my suggestion above is ok. It is not perfect, but its a solution causing minor problem in gerneral.
I am a bot programmer. And if i am writing software that interacts with another programm i must follow the rules defnied by use cases of foreign software. If the current index.php user interface makes it possible to not follow the software use cases then this is bad. The api.php interface is much more strict.
So i you like to have a bot that ignores rules definied by software then this it not possible. Currently a central interwiki data base is developed. I wish you good luck by talking to the developers to change the use case according to your rules. Bye. Merlissimo (talk) 20:30, 13 November 2011 (UTC)[reply]
Retrieved from "https://species.wikimedia.org/w/index.php?title=User_talk:MerlIwBot&oldid=1321442"