Template talk:Link with archives
Duplicate
[edit]This template duplicates existing functions of {{webarchive}}, with fewer features and a couple design problems. It should be redirected to {{webarchive}}, and use cases converted over. We don't want to have multiple templates doing the same thing, and then have a merge problem down the road.
Compare:
{{link with archives|http://www.nasa.gov/home/hqnews/2004/oct/HQ_04351_time_drags.html|As The World Turns, It Drags Space And Time|20041111092435|20060222051642|20080619060938}}
[http://www.nasa.gov/home/hqnews/2004/oct/HQ_04351_time_drags.html As The World Turns, It Drags Space And Time]. {{webarchive |url1=https://web.archive.org/web/20041111092435/http://www.nasa.gov/home/hqnews/2004/oct/HQ_04351_time_drags.html |date1=November 11, 2004 |url2=https://web.archive.org/web/20060222051642/http://www.nasa.gov/home/hqnews/2004/oct/HQ_04351_time_drags.html |title2=February 22, 2006 |date2=February 22, 2006 |url3=https://web.archive.org/web/20080619060938/http://www.nasa.gov/home/hqnews/2004/oct/HQ_04351_time_drags.html |title3=June 19, 2008 |date3=June 19, 2008}}- ↳ As The World Turns, It Drags Space And Time. Archived November 11, 2004, at the Wayback Machine. Additional archives: February 22, 2006, June 19, 2008.
The design issue is webarchive is WYSIWYG with explicit URLs defined, versus custom formatting in LWA that requires special code to unpack ie. one can not simply regex the wikitext for a URL. This is an issue for many bots and tools that depend on explicit URLs in the wikitext. Furthermore it encodes the original URL into a special template, rather than keeping it outside in a standard square link, this is also a design problem that will have impacts on many tools and bots. Furthermore it only seems to support the Wayback Machine and not the other dozen or so archive providers some of who don't use 14 digit time stamps. Furthermore it doesn't support different titles and source URLs like "Page 1", "Page 2", of a multi-page article. Finally, the reality is very few editors ever use this feature (multiple archives), it's very rarely used. For those who do need it, there is webarchive already in wide usage and supported by many tools and bots. -- GreenC 05:47, 19 May 2025 (UTC)
- Hi GreenC. It is indeed dedicated to the Wayback Machine (maybe another title will clarify this better). It follows the philosophy of “Do one thing, do it well”. “Less options” is also a goal, because it can guarantee a standard output and a less verbose input. You can also see the result by comparing in your examples the verbosity of {{webarchive}} (where you have to rewrite a lot of useless information) with that of this template. Consider also that the Wayback Machine is by far the most used archive tool on Wikipedia. As for the problems concerning bots, I have no idea. Originally I created it for Latin Wikipedia, where we basically don't use bots. --Grufo (talk) 06:39, 19 May 2025 (UTC)
- Yeah well we don't have multiple templates that do the same thing, and you are not addressing the issues I raised about why we do it the way we do it with so-called "verbosity". Obviously you have no intention of redirecting, so I will need to take it to templates for deletion as next step. -- GreenC 22:19, 19 May 2025 (UTC)
- Hi GreenC. I might not have understood why you value positively the verbosity of {{webarchive}}. From what I can read in your previous comments, you think that verbosity is a fair price to pay in order to have the possibility to support
- different titles instead of the archive dates
- other archive providers
- For me supporting #1 is a bad idea, because it favours the proliferation of non-standard outputs, while supporting #2 goes against the principle of having specialized templates that do only one thing. As for supporting multiple snapshots, it is only a side benefit made possible by the syntax of this template, it was never a main goal (although it is nice that it is possible). The initial goal was to be able to write a link accompanied by a 14-digit number without further verbosity. This also answers your last point about making this template a redirect: why would I ever create it in the first place if I did not think that something was partially wrong with {{webarchive}}? Once again, remember that this was originally created for Latin Wikipedia, where the choice was between importing {{webarchive}} verbatim or creating a new template from scratch. Once this template was created, I recognized its potential and I imported it to English Wikipedia as well. --Grufo (talk) 23:00, 19 May 2025 (UTC)
- Hiding archive URLs in the wikitext behind a coded number is a terrible mistake. It's no more a "price to pay" than
|url=in{{cite web}}. Keeping URLs in the clear where any tool can access them is essential for long term robustness and stability. In any case, your proposal here to create a custom format that hides the URLs will require the cooperation and buy-in of 1000s of tools and bots that need to be re-coded in order to support this custom URL hiding scheme, created without an RfC, consensus or discussion ie. it will never happen. Furthermore, the main bot that handles archives is Internet Archive Bot which runs on 300+ wikis and has global community approval, support from the WMF - it only supports webarchive. -- GreenC 00:24, 20 May 2025 (UTC)- I am not sure why you feel the urge to settle a dispute the day after this template was created. In my opinion it can coexist together with {{webarchive}} and cover a different audience (namely humans instead of bots). I am also pretty sure that you can give precious contributions to it. What I think is important though is that the outputs of both this template and {{webarchive}} get somehow harmonized (I have no opinions on the best output). I can give you some examples. On Latin Wikipedia we just created the {{la:Opus}} template, which is going to replace all the old {{la:Cite book}}, {{la:Cite web}}, etc. templates; in no way anyone is going to correct all the existing pages, but instead the two groups will simply continue to coexist… Until one will eventually fall into oblivion. The outputs they produce, however, is (almost) identical. In other cases we have templates that do almost the same thing but with slight differences and without one being the replacement of the other, and so they will keep coexisting happily. English Wikipedia also has tons of different template-link templates that do slightly different things, and that is also fine as long as the output produced is consistent. This particular template brings the benefit of brevity and readability in the input, which in a dense article with a tightly packed wikitext can make a big difference: what is the purpose of depriving English Wikipedia of this possible advantage? And in which way this is going to cause any trouble? My two cents. --Grufo (talk) 00:57, 20 May 2025 (UTC)
- First if all this is English Wikipedia not Latin so things here are different. You should look at the Templates for Deletion and see how frequently templates are deleted/merged when they serve the same basic functions.
- Your template is actually less friendly to humans, as it requires learning a new system developed without community feedback or input, based on positional arguments and coded numbers. It hides the URLs so all our standard tools can not see them: InternetArchiveBot, WaybackMedic, CitationBot, ReFill, AWB, VisualEditor, regex replacements, etc.. you have hidden the URLs from everything, now exposing them to link rot (archive URLs have link rot) making them more brittle and error prone. You have hidden them from WP:URLREQ so when a dead link can be made live again it will no longer work ie. dead links will stay dead forever even though they could be converted back to live links.
- Your template does not support basic functionality like changing how dates displayed (YMD, ISO,etc) which is a requirement since many articles require consistent date display. It does not support multiple web archives. It is no more "human friendly" than webarchive which is more in line with our standard CS1|2 template that everyone is already familiar with.
- Your standard of value here is simplistic: how many characters are used to invoke the template. For you that is all that matters. It is the programmers mindset, where the fewer letters the better, even if it means encoding and abstracting and increasing complexity. But we are not programmers we are humans. And simple things like
|url=and|title=and|date=can be understood by everyone it takes no thought and is self-documenting. And it's accessible to the widest audience including humans, semi-manual and fully automated editors with least amount of friction. -- GreenC 16:16, 20 May 2025 (UTC)
- I am not sure why you feel the urge to settle a dispute the day after this template was created. In my opinion it can coexist together with {{webarchive}} and cover a different audience (namely humans instead of bots). I am also pretty sure that you can give precious contributions to it. What I think is important though is that the outputs of both this template and {{webarchive}} get somehow harmonized (I have no opinions on the best output). I can give you some examples. On Latin Wikipedia we just created the {{la:Opus}} template, which is going to replace all the old {{la:Cite book}}, {{la:Cite web}}, etc. templates; in no way anyone is going to correct all the existing pages, but instead the two groups will simply continue to coexist… Until one will eventually fall into oblivion. The outputs they produce, however, is (almost) identical. In other cases we have templates that do almost the same thing but with slight differences and without one being the replacement of the other, and so they will keep coexisting happily. English Wikipedia also has tons of different template-link templates that do slightly different things, and that is also fine as long as the output produced is consistent. This particular template brings the benefit of brevity and readability in the input, which in a dense article with a tightly packed wikitext can make a big difference: what is the purpose of depriving English Wikipedia of this possible advantage? And in which way this is going to cause any trouble? My two cents. --Grufo (talk) 00:57, 20 May 2025 (UTC)
- Hiding archive URLs in the wikitext behind a coded number is a terrible mistake. It's no more a "price to pay" than
- Hi GreenC. I might not have understood why you value positively the verbosity of {{webarchive}}. From what I can read in your previous comments, you think that verbosity is a fair price to pay in order to have the possibility to support
- Yeah well we don't have multiple templates that do the same thing, and you are not addressing the issues I raised about why we do it the way we do it with so-called "verbosity". Obviously you have no intention of redirecting, so I will need to take it to templates for deletion as next step. -- GreenC 22:19, 19 May 2025 (UTC)
- There is already consensus to have a single template: Wikipedia:Templates_for_discussion/Log/2016_October_24#Template:Wayback. If/when I make a new TfD it would repeat the same thing: we need one universal template that is supportable by all humans and tools without needless fragmentation and the problems that causes. In the past 9 years Enwiki went from about 800,000 archive URLs to over 13 million, and more than 75% of that increase was due to bots, which was made possible by the standardization of templates, which are equally usable and accessible to bots and humans without problem. Webarchive is considerably less complex to use than
{{cite web}}and uses the same basic design features understood widely by the community, both humans and human tool makers, the argument made above that Webarchive is not human friendly is simply fallacious. Rather this new template represents one editor's personal preferences without regard for all the problems it causes system-wide. -- GreenC 16:42, 20 May 2025 (UTC)
- Hi GreenC. Personally, my criterion of what I value highly in a page is “what wikitext is easier to read” (not easier to insert/deploy – although this comes right after as a priority); the verbosity per se matters zero, as long as it is justified. Imagine a dense page containing your two initial examples: the one using {{webarchive}} can easily give a headache to an editor that approaches the page for the first time, while the other one is a bit less confusing to read. As for the other points you mentioned:
- Date formats: This is a feature that can be added easily (edit: added now)
- Dead links that come back to life: I did not know that we use to remove the link to the Wayback Machine in these cases. If I were to see the following link in a page, my instinct would be to leave it intact (because even though Politico.eu is a perfectly functioning website, if it had – hypothetically – gone dark once, it might go dark again in the future – furthermore an editor can also have had other reasons to add a link to the Wayback Machine, not necessarily a dead link):
- Archive URLs that rot: This is the first time I hear about this. How can a link to the Wayback Machine rot? Will it stop functioning? And how do we normally solve it? Do bots search themselves for an alternative snapshot in the Wayback Machine?
- Ablity to be parsed by bots: Let's imagine that all Wikipedia bots were able to parse the transclusions of this template; what would they do with them exactly? What would a bot ideally do with a link to Politico.eu like the previous one (which is built using this template)?
- You are right though that English Wikipedia is not Latin Wikipedia. I have extensively expressed my opinions and my concerns. Feel free to turn this template into a redirect if you think that is the best choice for English Wikipedia, I will not oppose it. But if you ask me to cast my vote, with the information I have, I vote against. --Grufo (talk) 00:39, 21 May 2025 (UTC)
- Hi GreenC. Personally, my criterion of what I value highly in a page is “what wikitext is easier to read” (not easier to insert/deploy – although this comes right after as a priority); the verbosity per se matters zero, as long as it is justified. Imagine a dense page containing your two initial examples: the one using {{webarchive}} can easily give a headache to an editor that approaches the page for the first time, while the other one is a bit less confusing to read. As for the other points you mentioned: