User talk:PDFbot/Archive 1

Problem with edits

Your edits are leaving a space between the size and closing parenthesis: Main Line (Pittsburgh to St. Louis) --NE2 23:01, 14 February 2007 (UTC)

PDFBot

In case it helps, you might look for the PDF code in standardize_notes.py, the RefBot code from four versions ago. (SEWilco 08:05, 15 February 2007 (UTC))

---

(Ed: 21:16, 5 March 2007 Taxman granted bot status to User:PDFbot)

PDFbot

PDFbot has attempted to specify the size of a PDF in this edit, but got it wrong, at 0.0 bytes. Could you look into the problem, please, and revert any similar instances. thanks --Tagishsimon (talk)

It looks like a one-off, having trawled some other PDFBot changes. Would be nice to know the cause & prevent a repetition, though. --Tagishsimon (talk)

I had also discover one instance where this happened, similar I had thought it was a fluke. Both the servers were running Apache/1.3.26 with ApacheJserv/1.1.2, but they report the values incorrectly when the bot queries them with the HTTP HEAD Request. I've since put in a few checks into the code that checks for a minimum size and MIME type. —Dispenser 08:27, 10 March 2007 (UTC)

PDF link

Where and why was it determined that the pdf link is problematic? It's never caused a problem on the pages that I've seen. Rlevse 01:11, 23 April 2007 (UTC)

The problem comes from a change to how the PDFlink CSS class works, it now gives all links the icon instead of just the external link. It is only apparent when the size of the links is correctly formatted using Binary units with links to their respective article (per MOS:NUM#Binary prefixes). Then there is the problem that this usage is not supported and makes machine parsing of the wikitext version a bit more problematic. If you feel that file size is useful enough to have, then propose at Template:Cite web for the addition of a size parameter. —Dispenser 01:38, 23 April 2007 (UTC)

Ah, okay.Rlevse 01:41, 23 April 2007 (UTC)

PDFbot error

I was looking through PDFbot's recent edits and I saw a mistake on PDFbot's part. See this edit. The problem is, PDFbot completely deleted the URL and did not place it into the PDFlink template as a parameter. « D. Trebbien (talk) 18:53, 9 May 2007 (UTC)

Here is a different edit that may be used to improve PDFbot in the future. For some reason, it fixed one PDF link, but not another in the next sentence. « D. Trebbien (talk) 19:06, 9 May 2007 (UTC)

Another error is exemplified with this edit. Somehow the first part of the PDFlink template got separated. Here is the text of the reference:

Kittipong Kittayarak, {{PDFlinkwith a simple majority could be elected in one constituency. The Senate was elected based on the province system, where one province can return more than one Senator depending onouioioo The court system (ศาล, saan) included a constitutional court with jurisdiction over the constitutionality of parliamentary acts, royal decrees, and political matters. The [[Thailand legislative election, 2001|319 KiB}}, National Taiwan University and Academia Sinica Asian Barometer Project Office Working Paper Series No. 28, 2004

PDFbot proceeds to update as if nothing is wrong. « D. Trebbien (talk) 19:16, 9 May 2007 (UTC)

I have found another PDFbot error ([1]). You should consider shutting down PDFbot. « D. Trebbien (talk) 20:23, 9 May 2007 (UTC)

I'll address each of the diffs

1. Unintended instance that was start by cron before the fix got in (see #4). Fixed cron entry so it only run on the first of the month, instead of trying the first Wednesday of the month.
2. (PDF) tag is referencing to the former. There should really be two there, one for each link. This was also part of the new code that I was testing.
3. State right in the documentation that it would replace the second parameter. Not to mention the wikitext there is malformed, I've added stricter tolerances to what can be interpreted as the template. (Ed: See also Oleg_Alexandrov: Is_there_a_tool_...)
4. Self identified, shutdown, identified and fixed the source, restarted.

I've also updated the online version since I've added a few things. —Dispenser 03:50, 10 May 2007 (UTC)

Hi Dispenser! I'm sorry, I had to shut down PDFbot after the above complaint. Please give me a shout when the bug is tracked down and I'll unblock (alternatively, others will of course be happy to do so as well!). Cheers  ⋐⋑ REDVEЯS 20:40, 9 May 2007 (UTC)

PDFbot in german wikipedia

I added the PDFlink template to the german wikipedia and noticed that the bot added the size to an templated link. Can I refer to this in my german documentation, as I do not know if this feature is intended. de:Benutzer:Roomsixhu 13.Mai. 2007 , 03:07 CET —The preceding unsigned comment was added by 87.187.45.132 (talk) 01:08, 13 May 2007 (UTC).

Sorry, that was wrong. Myself copied and pasted the size from an english wikipedia article. But now I am waiting if the bot will visit my PDFlink template in the artikle de:Differential_(Mathematik). --de:Benutzer:Roomsixhu —The preceding unsigned comment was added by 87.187.45.132 (talk) 02:00, 13 May 2007 (UTC).

I haven't written PDFbot to handle other languages yet; however, in the next few days I will look into running it on de: —Dispenser 04:24, 13. Mai 2007 (CEST)

[2] That was surprisingly easy, now to translate the interface and get approval.

Great, on de:Vorlage:PDFlink I have already tried to translate the Dokumentation, but still have not the categories there, finally the documention there for templates is on the discussion page, and think I made not too much mistakes using the termini technici. --de:Benutzer:Roomsixhu —The preceding unsigned comment was added by 87.187.11.153 (talk) 11:26, 13 May 2007 (UTC).

It already worked March 8. 2007: de:Spezial:Beiträge/PDFbot. Probably also a copy and paste link from here. --de:Benutzer:Roomsixhu

Unfortunately the template PDFLink has been deleted in german Wikipedia. Cheers de:Benutzer:Roomsixhu

PDFbot

Your bot PDFbot has now botstauts at German Wikipedia! -- sk 18:40, 22 May 2007 (UTC)

Template:PDFlink

Hi, could you kindly explain the mysteries of Template:PDFlink? Where is the missing icon and what is the meaning of {{{1}}}? Thank you. RCSB 12:06, 28 May 2007 (UTC)

{{PDFlink}} works differently from its previous incarnations which was an image and link to the PDF page. But now, the template places a span wrapper around the first unnamed parameter ({{{1}}}) and has a second parameters for the file size. The span wrapper has class="PDFlink" (see MediaWiki:common.css) which overrides the background image on the hyper-link. This feature, however, is only useful for Internet Explorer 5-6 Users as newer browsers support the advanced CSS selectors that adds the icon to external links containing .pdf.

I also run a bot that attempts to convert the old format to the new format, while adding file size information. The bot runs four times a week on Category:PDFlink without a parameter and once a month on all transclusions of PDFlink. —Dispenser 04:05, 29 May 2007 (UTC)

Thanks for the explanation. Impressive. RCSB 15:10, 29 May 2007 (UTC)

PDFbot binary prefixes

I hate to bring this up, but on what basis does PDFbot add {{PDFlink}} tags that use IEC60027 binary prefixes (KiB), etc.? It seems like they are highly disputed, and it would be best not to contribute to the dispute by doing so. I was also under the impression that, while there is no concensus, using both KB and KiB seemed to have a lot more support than simply using KiB. If I missed a previous discussion about this in the context of PDFbot, please accept my apologies (in any event, please accept my apologies for bringing up this issue at all). jhawkinson 09:18, 1 June 2007 (UTC)

This issue was covered in the requests for approval. It is completely unambiguous. The other option of course would have been to actually make it base-10, but then there be more confusion as it wouldn't have matched up with what the OS would telling them.

KB is absolutely not ambiguous. First K is not a decimal prefix, second files size are never reported in decimal unit, not on Unix, not on OS X, not on Windows, not on z/OS, hence the 'other' option would have been indeed to present what the OS would be telling them : xxx KB -- Shmget 06:29, 5 June 2007 (UTC)

Thus far the only problems have been editors copying the syntax elsewhere from the page and not filling in the byte size in comment correctly but updating the documentation has seemly stop that. —Dispenser 00:15, 2 June 2007 (UTC)

Indeed. Are you aware the situation has changed somewhat? You might browse Wikipedia talk:Manual of Style (dates and numbers)#Binary prefixes. It seems there is quite a lot of dispute. My concern is that PDFbot has chosen a side in this debate, and is actively taking actions which promote that side at the expense of the other (of course, PDFbot does other--very useful--work!). For instance, in part due to PDFbot, the proponents of IEC 60027 prefixes can claim that number of Wikipedia pages using IEC60027 prefixes is increasing daily at some rate. But PDFbot's contributions to that do not represent editors deciding and accepting the new prefixes.

You list two options above, but of course, you omit the return to ambiguity, and I think that's an important one. For the vast majority of uses (especially in the context of PDFbot!), the distinction of 8% is not material -- it does not significantly affect download estimates; also, anyone who is likely to care should understand from context what is meant. In my mind, the biggest problem with the IEC60027 notation is that it is confusing to new users, and most of PDFbot's references will be seen by those unfamiliar with the notation, and we should avoid throwing foreign notation at them for this reason. (I also personally find the notation cumbersome). I'm not sure what to suggest. This is still highly in flux over in MOS, but one answer is the compromise solution of listing both. While I still find it cumbersome, I do think it would be preferable to the current situation of only providing the IEC60027 prefix that is foreign to most readers. Thanks for your time. jhawkinson 12:25, 2 June 2007 (UTC)

In any case, if I change inserted style it would mean roughly 2,000 pages would have to be re-saved. I'm not going to do this until the debate is settled at MoS, but if they want to use the "what links here" count then they should only look at the redirect pages (KiB)

Not at all! The bot can simply ignore the units specified on the current page when deciding if the page needs an update, and only use a compromise (or whatever) change going forward. I fear it will be some time before MoS gets any more settled than the current "there is a lot of dispute" state. Anyhow, I wanted to bring this to your attention, and I think I've done so. Thanks! jhawkinson 14:40, 2 June 2007 (UTC)

Virtually no one in the computer industry or publishing industry uses the IEC60027 binary prefixes. A group of people on Wikipedia are evangelizing this peculiar form of binary units. The original proposal on the Manual of Style was "The MoS should encourage the use of the IEC prefixes in all binary-multiple contexts". It is Wikipedia's job to encourage the use of a standard virtually no one uses?

I recall reading on a PDFbot related talk page that it is good to expose more readers to the IEC prefixes. If they see a file is 150 KiB they can click on KiB and learn the Truth about binary units. The IEC prefixes are unambiguous and more accurate. Most reader just what to know if the PFD file is 100 kilobytes or 4 megabytes.

It more people don't start using the IEC binary prefixes, Wikipedia with have editors ringing door bells to tell the world about the benefits of kibibytes and mebibytes.

These comments are just a little over the top but I think Wikipedia should use what the world uses, not try to fix the world. I also think PDFbot is a useful addition to PDF links. -- SWTPC6800 04:25, 5 June 2007 (UTC)

PDFbot with Cite web

Is it possible to adapt this BOT to add filesizes to citations based on {{cite web}}?

• e.g. {{cite web |url=<url> |title=<title> |format=PDF |...}}

→ AA ^{(talk • contribs)} — 12:26, 11 June 2007 (UTC)

Thus far nobody besides myself has pushed for a |size= parameter in web cite. And I'm not going to start work until the binary vs. SI prefixes dispute is resolved. Other than those two thing I'd like to work this feature. —Dispenser 16:48, 11 June 2007 (UTC)

Thanks for the heads-up. I think this would be a useful addition to the template. → AA ^{(talk • contribs)} — 21:53, 11 June 2007 (UTC)