The {{Cite_web}} citation template is currently fairly cumbersome, and contains parameters that seem unnecessary or are confusing in their usage. It could do with an update to simplify it, making it easier to use, and bring it inline with other citation templates, in regard to the handling of archive links.
Contents
Proposed updates
- For starters, the archiveurl and archivedate parameters must both be populated otherwise it displays an error message. This could be changed so only one is needed, much like other citation templates do. Indeed, its technically possible to build the web.archive.org url using just the archivedate value. By providing only one value, such as archivedate, this also shortens the template and keeps the bytes down on each article.
- Another tweak revolves around author-related parameters, such as last, first, and coauthors. These last three seem unnecessary, and the author parameter can handle all that by itself, much like {{BookCite}} does. The usage would read: Name(s) of the author(s). Must be formatted like "Last name, First name"; multiple authors must be separated with a ";" between their names. Note that an author is not necessarily an individual and could be an organization. A long list of authors can be reduced using ", et al." after the first name. Provide wikilinks whenever they are available.
- Much like other existing citation templates, additional backup messages can be added, under various conditions such as when nobackup=1 is used or when archiveurl/archivedate have not been provided:
- (backup link not available)
- (backup link not verified!)
- Minor tweaks to the error message when no url or title are provided, these can be split so the message tells you exactly what is missing rather than say both fields are needed, so we have a unique message for each condition:
Missing url parameter on {{Cite_web}}: Parameters url and title must both be specified.Missing title parameter on {{Cite_web}}: Parameters url and title must both be specified.
- UPDATE: Shortened the error message below.
- Cite_web error: missing url parameter must be specified.
- Cite_web error: missing title parameter must be specified.
- Other tweaks, to bridge the gap left behind by the above changes, such as when the archivedate is provided, but the archiveurl isn't, it still provides the "Archived from the original." text with the original linked url. But of course, without a date that it was archived.
Further updates based on discussion
- UPDATE 1: Use url to automatically display the domain name, if the work parameter is empty. Previously, you would have to manually enter the website name in one of the work or publisher parameters. This is unnecessary when it can be automated.
- UPDATE 2: Add a nolive parameter, to flag when the original online content is no longer available, regardless of an archive link. When nolive=1 is used, the following is appended:
- (content now obsolete)
- UPDATE 3: Anchor the position of the date so it appears in a single position rather than moving around based on whether or not the author is present. I've placed it after between title and work.
- UPDATE 4: Remove all reference to archiveyear in the code, given the widespread use of archiveurl or archivedate, and the lack of mention within the template documentation.
I have created a prototype version of the current Cite_web template, {{Cite_web2}}, and modified it with the above changes, which I have so far tested. Please checkout the case examples below.
Case examples
archiveurl & archivedate:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|date=2021-04-28|archiveurl=https://web.archive.org/web/20080517034311/https://www.miniaturetrading.com/im/selectCard/series1025|archivedate=20080517034311}}
Series 1025 on www.miniaturetrading.com (April 28, 2021) (backup link archived on May 17, 2008)
archivedate only + authors:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|author=[[Ben Solo|Solo, Ben (Kylo Ren)]]; [[Han Solo|Solo, Han]]|date=2021-04-28|archivedate=20080517034311}}
Series 1025 by Solo, Ben (Kylo Ren); Solo, Han on www.miniaturetrading.com (April 28, 2021) (backup link archived on May 17, 2008)
archivedate only, no author:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|work=[[Star Wars Miniatures]]archivedate=20080517034311}}
Series 1025 on Star Wars Miniatures (backup link archived on May 17, 2008)
no work, so domain automatically added:
*{{Cite_web2|url=https://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|archivedate=20080517034311}}
Series 1025 on www.miniaturetrading.com (backup link archived on May 17, 2008)
archiveurl only:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|archiveurl=https://web.archive.org/web/20080517034311/https://www.miniaturetrading.com/im/selectCard/series1025}}
Series 1025 on www.miniaturetrading.com (backup link archived on May 17, 2008)
language and format:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|language=German|format=PDF|archivedate=20080517034311}}
Series 1025 (German) (PDF) on www.miniaturetrading.com (backup link archived on May 17, 2008)
no archive links:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025}}
nobackup=1:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|nobackup=1}}
Series 1025 on www.miniaturetrading.com (content obsolete and backup link not available)
Legacy last, first, coauthor - demonstrates removed params don't break the template:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|author=[[Ben Solo|Solo, Ben (Kylo Ren)]]; [[Han Solo|Solo, Han]]|last=Skywalker|first=Luke|coauthor=Darth Vader|archivedate=20080517034311}}
Series 1025 by Solo, Ben (Kylo Ren); Solo, Han on www.miniaturetrading.com (backup link archived on May 17, 2008)
Error: no url:
*{{Cite_web2|url=|title=Series 1025|archivedate=20080517034311}}
- Template:WebCite error: missing url parameter must be specified.
Error: no title:
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=|archivedate=20080517034311}}
nolive=1 and archive links
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|nolive=1|archiveurl=https://web.archive.org/web/20080517034311/https://www.miniaturetrading.com/im/selectCard/series1025|archivedate=20080517034311}}
Series 1025 on www.miniaturetrading.com (original page now obsolete)
nolive=1 and nobackup=1
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|nolive=1|nobackup=1}}
Series 1025 on www.miniaturetrading.com (content obsolete and backup link not available)
nolive=1 but no archive links
*{{Cite_web2|url=http://www.miniaturetrading.com/im/selectCard/series1025|title=Series 1025|nolive=1}}
Series 1025 on www.miniaturetrading.com (content now obsolete; backup link not verified!)
Potential other updates
The following are optional parameters, but removing or changing them could help simplify the template:
- Remove accessdate, we don't capture this information on other citation templates.
- Remove year and month parameters, as date could handle this information.
- Maybe, combine work and publisher parameters into a single parameter. For example, the value could be input as "website (publisher)". UPDATE: one option is keep the publisher parameter and remove work; or remove one and rename the remaining one to something thats clearer to the user what its asking for, such as source maybe.
If anyone has any other opinions or ideas on what else could be tweaked, feel free to add your thoughts below. Plume Tray (talk) 20:19, 26 April 2021 (UTC)
Implementation
There will be two kinds of changes necessary, the template change itself and a bot run on all existing instances to remove unwanted parameters as described above. The template change is the simplest and only requires the source code from {{Cite_web2}} copied over to {{Cite_web}}. Once saved, it will not break any existing instances as the new updates are compatible. Next, a bot script has been prepared to carry out the following changes to existing instances:
- Remove last, first, coauthors, accessdate, year and month.
- If author is empty but last and first are not, use those values in author.
- Remove archiveurl if it has a web.archive.org link, and use its date in archivedate if its also empty. If the web.archive.org link after the datestamp is different from url, archiveurl will be kept and archivedate deleted instead. Instances using archive.today will be kept.
Discussion
- Fantastic work, Plume Tray! I like all of the proposed updates! As for the potential other updates, I would disagree with removing the accessdate only because that's standard for online citations in multiple style guides (ex. MLA and APA) and may differ from the date of publication and the date of the archive. Is it useful to know? Ehhh. But it's one of those things that's typically done. Date parameter, yeah, agree with simplifying that and people can just format it correctly themselves. Work and publisher, I honestly get confused all the time on what to put anyway, "wait is this website a work or a publisher?" So anything to make that less muddy is good with me. Immi Thrax
(talk) 20:54, 26 April 2021 (UTC)
- Thanks Immi, we don't seem to include accessdate on many of our other citation templates so it seemed inconsistent to keep it here. Work/publisher are a strange pair because quite often one or the other is just populated with the domain name thats already in the url param, so why duplicate - instead the template could do some processing on the url to automate retrieval of the domain name to display without the need to input manually. I'll have a look through existing examples of the work/publisher params in action, so I can better understand its range of uses across Wookieepedia. Plume Tray (talk) 23:32, 26 April 2021 (UTC)
- Further to my reply, I've gone ahead and added the ability to automatically display the website name when the work param is empty in the template. I've added this item to the Proprosed updates section above too. Plume Tray (talk) 08:36, 27 April 2021 (UTC)
- Thanks--I'll ponder it some more myself. Something I thought of that's bugged me when citing offline sites that still exist in archive form is how do we indicate that?! Maybe something like a nolive=1 parameter similar to nobackup? Immi Thrax
(talk) 04:58, 28 April 2021 (UTC)
- A nolive=1 is simple enough to add. I wonder if an archive version of Cite_web template would be an option, much like with {{SW}} and {{SWArchive}}. The upside to this is just change the name of the template used in an article eg Cite_web_archive, without adding/changing params, however the downside is maintaining another complex template. So maybe nolive=1 would just be easier to maintain in the long term. Certainly having a param like this could also produce an inline message on the end of the template saying (content obsolete) or (content unavailable). Plume Tray (talk) 18:06, 28 April 2021 (UTC)
- I like this idea. We already use "content obsolete" on templates where the website is defunct, but that's something I would also love to see on Cite_web AND elsewhere, and this parameter fit the bill perfectly. -- NanoLuukeCloning facility 21:58, 28 April 2021 (UTC)
- I've added the nolive parameter to the prototype template and now included it in the Proposed updates and examples sections above. Plume Tray (talk) 23:45, 28 April 2021 (UTC)
- Seems good! -- NanoLuukeCloning facility 09:11, 29 April 2021 (UTC)
- A nolive=1 is simple enough to add. I wonder if an archive version of Cite_web template would be an option, much like with {{SW}} and {{SWArchive}}. The upside to this is just change the name of the template used in an article eg Cite_web_archive, without adding/changing params, however the downside is maintaining another complex template. So maybe nolive=1 would just be easier to maintain in the long term. Certainly having a param like this could also produce an inline message on the end of the template saying (content obsolete) or (content unavailable). Plume Tray (talk) 18:06, 28 April 2021 (UTC)
- Thanks--I'll ponder it some more myself. Something I thought of that's bugged me when citing offline sites that still exist in archive form is how do we indicate that?! Maybe something like a nolive=1 parameter similar to nobackup? Immi Thrax
- Perhaps a further tweak to the error message, to shorten it a bit to something like: Cite_web error: missing url parameter must be specified. Plume Tray (talk) 23:32, 26 April 2021 (UTC)
- I've updated the error message in the template and the Proposed updates section above. Plume Tray (talk) 08:36, 27 April 2021 (UTC)
- First, I must says I'm glad you decided to tackle this one ^^. Second, I agree with the proposed updates. I would like to read ecks take on "archivedate/archiveurl", as I remember (maybe incorrectly) that he identified possible problems with the change. Regarding "author", given I authored BookCite, I can only agree ^^. For "work", the automation seems solid. I would not be against the removal of "accessdate"; in a normal citation it's meant to says: I accessed the version of this page at this day, but it's highly redundant with "archivedate", as it will always be the archived version that will be relevant to the citation (I had to cite a page twice recently to show the difference between versions). I'm also all for simplifying "date" (much like for BookCite), it's really overcomplicated in the live version of Cite_web; the template could do with the date changing place when no author is attributed. Finally, I don't think it's useful to change "work/publisher". A lot of people, including myself in the first half of the missing archive link project, have been wrongfully using publisher instead of work. But they are not meant to be interchangeable. Work is the website, while publisher is the owner of the website. Example: |work=Kotaku|publisher=G/O Media. Thanks for bringing this to the Senate Hall like I suggested, Plume. Given how wildly spread the use of this template is, this might require the input of the administrators, as well as status article reviewers, and even a Consensus Track vote. But it's moving forward ;) . -- NanoLuukeCloning facility 16:32, 27 April 2021 (UTC)
- Thanks Nano. Regarding "date changing place when no author is attributed", currently it displays between author and title when author is attributed, but when its not, it shows up after title and work. I've added dates to the examples above for clarity. What's your proposal? Plume Tray (talk) 18:06, 28 April 2021 (UTC)
- Just realized that I wrote "the template could do with the date changing place when no author is attributed" instead of "without". So what I meant is that the template could really win in terms of internal readability if we removed altogether the change of position of "date", to keep it always before "work". Also, I forgot to add this: it will need a solid bot work, as you're well aware. It's not yet the time to worry about that, but when we do commit to changes, we need 1) to be certain "this is the final version" (I really insist on taking all the time we need here), and 2) to have a solid workflow for bot run(s), with a selective test run at first. Anyhow, I hope this get some more input in the coming days/weeks ^^. -- NanoLuukeCloning facility 21:58, 28 April 2021 (UTC)
- I think a fixed date position is a good idea. But note that if no author, the date will show up as the first inline item, before the title. Would it be better if we placed it in a fixed position immediately after the title? Plume Tray (talk) 00:07, 29 April 2021 (UTC)
- IDK, honestly. One way or the other, I don't minde. As long as we remove the bunch of code lines handling the unnecessary position change, I'm happy ^^. -- NanoLuukeCloning facility 09:11, 29 April 2021 (UTC)
- I think a fixed date position is a good idea. But note that if no author, the date will show up as the first inline item, before the title. Would it be better if we placed it in a fixed position immediately after the title? Plume Tray (talk) 00:07, 29 April 2021 (UTC)
- Just realized that I wrote "the template could do with the date changing place when no author is attributed" instead of "without". So what I meant is that the template could really win in terms of internal readability if we removed altogether the change of position of "date", to keep it always before "work". Also, I forgot to add this: it will need a solid bot work, as you're well aware. It's not yet the time to worry about that, but when we do commit to changes, we need 1) to be certain "this is the final version" (I really insist on taking all the time we need here), and 2) to have a solid workflow for bot run(s), with a selective test run at first. Anyhow, I hope this get some more input in the coming days/weeks ^^. -- NanoLuukeCloning facility 21:58, 28 April 2021 (UTC)
- Nano, regarding combining work and publisher parameters, I meant in the same vein as the author or date parameters, where a single parameter would serve a dual purpose, for example, you can input "website.com (website owner)" and that way the average user can't go wrong trying to remember which info goes in which parameter, just put in one place. Plume Tray (talk) 00:07, 29 April 2021 (UTC)
- You're right, we don't really need an additional parameters as it doesn't have any interest in term of formatting. -- NanoLuukeCloning facility 09:11, 29 April 2021 (UTC)
- Thanks Nano. Regarding "date changing place when no author is attributed", currently it displays between author and title when author is attributed, but when its not, it shows up after title and work. I've added dates to the examples above for clarity. What's your proposal? Plume Tray (talk) 18:06, 28 April 2021 (UTC)
- Good job on modernizing this template. I agree that we don't need both archivedate and archiveurl since archivedate can be parsed to form both a legible date and a timestamp for Internet Archive. I also agree with the other proposed updates. Accessdate is, like you say, unnecessary and since one is required to provide an archive link upon adding an external link, that kind of serves as an access date. 1358 (Talk) 21:00, 6 May 2021 (UTC)
- Thanks Ecks. Plume Tray (talk) 11:47, 8 May 2021 (UTC)
- Pretty much per Ecks. This is great stuff and will be a very good update to such an important template. Very nice work with this. Tommy-Macaroni 17:06, 8 May 2021 (UTC)
- Extremely great stuff! And chiming back in since I brought up accessdate that I now agree it's not needed; the purpose it serves in other style guides is covered by our use of archivedate. This update is gonna be awesome. Immi Thrax
(talk) 02:06, 9 May 2021 (UTC)
- Extremely great stuff! And chiming back in since I brought up accessdate that I now agree it's not needed; the purpose it serves in other style guides is covered by our use of archivedate. This update is gonna be awesome. Immi Thrax
- Pretty much per Ecks. This is great stuff and will be a very good update to such an important template. Very nice work with this. Tommy-Macaroni 17:06, 8 May 2021 (UTC)
- Thanks Ecks. Plume Tray (talk) 11:47, 8 May 2021 (UTC)
- Note for when (since it's seems more and more that it's going to happen) you create a workflow to bot this mess: use the archiveurl to systematically check the url= and change it accordingly, because from experience, I know a certain number of Cite_web url and archiveurl don't match, and it will break archivelink on the Cite_web 2.0 if not corrected. -- NanoLuukeCloning facility 22:49, 6 May 2021 (UTC)
- That's a valid point, and I have seen this myself, so I'll make sure to add it the workflow. Though I do have to question why there are differences, and should the url value be updated to match archiveurl or vice versa? Plume Tray (talk) 11:47, 8 May 2021 (UTC)
- It happen because of link rot and redirects. A lot of Cite_web (and naked links) went without archive for a long time. Url changes, and redirects will result in editors like myself to open the url from an article, with the opened page been in fact redirected without the editor noticing, so when saved the archiveurl will be different than the url. I know I did it more than a few time, first because I was unaware of the issue, and later because I just missed it. I try to be indful of it nowaday. So, to make it clear, url= should be updated to match archiveurl part after https://web.archive.org/web/tiemstamp/. -- NanoLuukeCloning facility 13:00, 8 May 2021 (UTC)
- Ah yes, that makes sense now. I have a bot script that's 99% ready, and currently it will check if there is a match between url and archiveurl, and therefore keep archiveurl if they are different. I'm in two minds about updating url from that part after the timestamp in archiveurl, because it could run into a number of different issues that I may not be able to fully account for. For example, a perfectly working url, may get overridden by the url in archiveurl that might not be live anymore and not correctly redirect, and the bot will not know without some deeper form of URL checking, which may or may not work on websites that aren't bot friendly. Its safer just to leave that bit untouched for now, perhaps revisit that at a later date (I'll keep it in mind though and carry out some tests). Plume Tray (talk) 16:29, 8 May 2021 (UTC)
- It happen because of link rot and redirects. A lot of Cite_web (and naked links) went without archive for a long time. Url changes, and redirects will result in editors like myself to open the url from an article, with the opened page been in fact redirected without the editor noticing, so when saved the archiveurl will be different than the url. I know I did it more than a few time, first because I was unaware of the issue, and later because I just missed it. I try to be indful of it nowaday. So, to make it clear, url= should be updated to match archiveurl part after https://web.archive.org/web/tiemstamp/. -- NanoLuukeCloning facility 13:00, 8 May 2021 (UTC)
- That's a valid point, and I have seen this myself, so I'll make sure to add it the workflow. Though I do have to question why there are differences, and should the url value be updated to match archiveurl or vice versa? Plume Tray (talk) 11:47, 8 May 2021 (UTC)
- Like other citation templates, the archive feature should be just like the following. - JMAS
Hey, it's me! 11:53, 12 May 2021 (UTC)
| Instructions for providing a backup link
In accordance with our Sourcing policy, all links to external sites must include a link to a backup version, preferably using Internet Archive. Instances that do not include one of the following parameters will result in the page being placed in the maintenance category. Please only use one of the following parameters.
Note that for citation templates using a central repository, archivedate and archiveurl values must be added to the repository, and not on the individual instance of those templates. If this template uses one, it is located at Module:ArchiveAccess/SH:Cite web template update. |