User:Jakob.scholbach/zeteo/bugs

Comment by Jitse Niesen

(copied from Wikipedia talk page of Math project)

1. If it doesn't recognize the author, it asks the user to check. That is good. However, there are different spellings of a name ("Arnold" and "Arnol'd", to give a simple example), sometimes wildly different. I think it would be good if you could browse the author names in the database alphabetically. Does the system warn you if a similar name exists (like it does for publishers)?
2. Searching for "Arnol'd" gives some errors. You should be very careful with user-supplied data. See for instance the PHP manual for details.
3. Search for "Golub" indeed gives one of the books that I imported. However, the reply is {{Citation | last1=Golub | first1=Gene H. | author1-link=Gene H. Golub | last2=Van Loan | first2=Charles F. | author2-link=Charles F. Van Loan | title=[[:en:Gene H. Golub|Matrix Computations]] | publisher=Johns Hopkins | edition=3 | isbn=978-0-8018-5414-9 | year=1996}} The link in the title field is wrong.
4. Searching for "golub" does not yield any results; apparently the search is sensitive to case. I would make it case-insensitive.
5. I find the list of references that you get when you go to http://zeteo.info/ rather confusing. Start with an empty display.

It's a good start, but we'll have to see whether it will prove to be useful. But I've bookmarked the page.

I don't know of anybody in maths using EndNote, except for students who don't know latex and write their reports in Word. -- Jitse Niesen (talk) 03:07, 6 August 2007 (UTC)

3.-5. are fixed. 2. is partly fixed (I need to test it further)

Ad 1: When adding a reference, the query for the author is sharp, i.e. not allowing merely similar names. There is no warning if a similar author exists. Both is on purpose (otherwise, as the database grows, it will likely yield unwanted results). If it happens that someone adds a reference with author="Arnol'd" and another one with "Arnold", it is possible to redirect one author to another one (similarly with publishers and journals). This way you can handle different transcriptions, typing errors etc. After setting the redirect, only the "correct" name (of the author which gets the redirect) will be shown. As for looking up the authors alphabetically, you can do that in the author's view, sort it by name by clicking at the "Name" column. Jakob.scholbach 05:04, 6 August 2007 (UTC)

ISBN database

How is the interface with Diberri's tool supposed to work? http://zeteo.info/references.php?action=lookUpISBN&isbn=0120471418 does not recognize its output for some reason. -- Jitse Niesen (talk) 03:25, 29 August 2007 (UTC)

And another thing. I went to http://zeteo.info/references.php?action=presave&id=552, clicked on "edit this item", changed the auther field to "Sencar, Husrev T.; Ramkumar, Mahalingam; Akansu, Ali N." and pressed save, but it does not seem to work; the resulting page contains "2004array(1) { [0]=> string(4) "2004" } NULL NULL", so I guess there is a bug somewhere. By the way, I hope that you do make sure that the database is backed up regularly? I can store a copy on my computer if you want. -- Jitse Niesen (talk) 03:42, 29 August 2007 (UTC)

I've fixed the bugs you mentionned.

As for the ISBN handling: You can input an ISBN in the look-up-field in references.php. If nothing is found, it will show a link which will then look up the appropriate page of the Diberri tool. If something is found by the tool (it looks up an ISBN database), the result is shown in the input boxes and you can edit further. Then click at "Save" etc.

So, thanks again for the testing. I myself don't have that much time, so I'm grateful for comments. Also, if you feel that there should be other features, please be sure to let me know. I'm waiting for User:CBM's help with pulling out all the existing references of the Math articles, then I'll try to incorporate them into the database. I may need some help with this. Jakob.scholbach 16:34, 2 September 2007 (UTC)

As for the storage of the DB: thanks, it is stored at a good hosting agency. Jakob.scholbach 16:40, 2 September 2007 (UTC)

Well done. More requests follow ;) Firstly, it would be nice if page ranges would automatically use an en dash. So, if you input "30-45" or "30--45" in the Pages field then the program should change it to "30–45". And a cosmetic thing: If you search for "Akritas", then you get the same reference twice, probably because it has two authors with that surname. -- Jitse Niesen (talk) 01:30, 3 September 2007 (UTC)

The page formatting is done (however, I chose not to change the user's input directly (to avoid confusion on the user's part), but only to adapt it when it is exported to the Citation template). Jakob.scholbach 21:36, 3 September 2007 (UTC)

The other bug is fixed, too. (Took me some time to figure out how to write the query...) Jakob.scholbach 20:26, 11 September 2007 (UTC)

First name/last name

When you enter a reference, the page has a footnote at the author field which says

expected format: name // first name name // name, first name. separate several authors by ";"

However, when I enter

Ingemar Bengtsson; Åsa Ericsson; Marek Kuś; Wojciech Tadej; Karol Życzkowski

it does not recognize that Ingemar is the first name and Bengtsson is the family name. What's going wrong? -- Jitse Niesen (talk) 08:33, 16 September 2007 (UTC)

Thanks. Actually I was stumbling over this bug, yesterday, too. It is now fixed. (If a token consists of n words (not separated by a semicolon), the first n-1 words are taken to be the first name, the last one to be the name). If you have any comments on how to ensure a high quality of the database content, please have a look at the discussion at my talk page. Jakob.scholbach 15:31, 16 September 2007 (UTC)

Edition

It seems not possible to set the "edition" field to "3rd"; the program always truncated it to "3". -- Jitse Niesen (talk) 06:44, 21 September 2007 (UTC)

Thanks again for the testing. This was half of a bug. I now chose the following: the database internally saves only the number, i.e. if you input "3rd", it will store "3". Whenever an output is generated, however, the "3" will be made to "3rd" etc. This seems to be better than storing the string "3rd" in the db, because users in other languages will want to see "3." (German) or "3ième" (French) etc. Jakob.scholbach 21:49, 21 September 2007 (UTC)

Some issues

• This one [1] is the "third millenium" edition, a pun on it being the 3d edition of the book and being published in 2003. I was unable to add the edition in the database.

• I'd like to merge that one with this one [2] but I don't know how.

• This one [3] was reprinted by Dover in 2002. I like to include that fact in articles but I can't see where to type it into the database.

— Carl (CBM · talk) 20:45, 27 November 2007 (UTC)

OK, thanks. The first is kind of a bug. I only allowed for integers in the edition field, but I will change this. Ad 2: at the moment it is not possible to delete references, because I was afraid of a "deleting vandal". Nor is it possible to redirect one reference item to another one. So, at this time, the information will be there twice. Ad 3: what would the output in terms of the reference like? Something like

"{{citation...}}, reprinted by Dover in 2002"?

I currently have a "comment" field, which has no function yet. I could imagine to use this field and just append this to the template. But actually I think in this particular case it is best to create a second item for the reprinted edition (it probably also has a different ISBN, so it really is a different book in a sense). Jakob.scholbach (talk) 21:25, 28 November 2007 (UTC)

For the assembling different editions of a book into one item or relating them: I guess I will introduce some mechanism which allows this. This can also be useful for translations of a book etc. Then the database at least shows you that there is a reprint, and you would then have to add the information of the reprint edition manually to WP articles. OK? Jakob.scholbach 11:54, 1 December 2007 (UTC)

The edition bug is now fixed. You can now input strings for the edition, too. Jakob.scholbach 14:38, 1 December 2007 (UTC)

Thanks. I think that adding the reprint information manually is fine, and it would allow the database to have full info about both versions, which makes sense. By the way, we are using zeteo a lot in rewriting mathematical logic. If I import a citation for a book that is already in the database, will the system know to merge them, or will it make a duplicate? I am planning to import the new refs from that article once it's done. — Carl (CBM · talk) 16:41, 1 December 2007 (UTC)

OK. The DB will warn you whenever an existing reference is pretty similar to the one you want to add. Then, you can decide whether you want to add it anyway or not. At this time, it does not really merge the information of the new one with the data of the existing one. For example if an existing entry does not have a URL, but the (similar) one you are importing, it is not clever enough to update the url without further manual effort. Jakob.scholbach 17:14, 1 December 2007 (UTC)

Please use my talk page for further bug reports etc.

Thanks. Jakob.scholbach 17:16, 1 December 2007 (UTC)