Template talk:Firefox release compatibility

New 'ISA' layout
In the contrib template, shifting the ISA info into a separate column allows for the final releases per OS arch to be presented without excessively long strings in the 'Latest version' column, e.g the final SPARC builds for OpenBSD. It also avoids the need to specify 'Intel' or 'PPC' regarding two macOS versions in the Operating system column. The new layout is more consistent.

Most of the information concerning unsupported Firefox builds were not useful as they were targeted at OS releases that have reached EOL and are now unsafe to use online. The new design puts the current releases centre-stage, the reader can infer that older versions of the various platforms are unsupported. If there is a method of hiding rows within a template, then this information could be restored from old revisions unobstrusively.Paianni (talk) 16:24, 7 December 2017 (UTC)
 * Hi.
 * Actually, what you did is quite a common mistake, but a mistake nonetheless. Wikipedia is an encyclopedia and values history.
 * That said, the table that you created is totally meaningless. For example, the column for Windows NT, x64 says: "57.0.1 (6.1 and later) and 52.5.0esr (6.1 and later)". What on earth is "6.1 and later"?
 * Best regards, Codename Lisa (talk) 16:39, 7 December 2017 (UTC)
 * I used the NT version numbers so I could avoid listing all the marketing names for the server and consumer releases of Windows, and to encompass derivative releases which Firefox would run on with no issues. 6.1 is analogous to 7/Server 2008 R2, 5.1 is analogous to XP. Partially due to this I could condense the information for current Windows releases into two rows, whereby the previous revision uses three and duplicated URLs between rows. Paianni (talk) 19:01, 7 December 2017 (UTC)
 * Hey, guys. I just took a peak at the disputed revision: . Quite frankly,, do you have the slightest idea what you are doing? You are defeating the purpose of this template. Its purpose is exactly showing those things that you deleted. (Hello? Did you read the template title?) The things that you don't delete are already in the infobox, and later in the system requirements table, i.e. the latest version number and the OS requirement.
 * Of course, I can see that you have pulled the same stunt with mobile versions and unofficial ports too. Nobody has contested those so far. That does necessarily mean it is okay; it could also mean that nobody read them, nobody understood them, or nobody cares about them.  Fleet  Command ( Speak your mind! ) 13:17, 8 December 2017 (UTC)
 * My proposed layout is a double-edged sword. On the one hand, dividing up the ISAs allows for the last releases per OS ISA to be presented in a sensical way, the contrib template I conceived actually has more info now than would have made sense with the old layout. On the other hand, information about older releases not forming the final release of a particular ISA don't have a logical place to go. My interpretation of the table's purpose is direct links to the servers hosting the installation files, which would only be sensible for current OS releases as the template should not encourage use of unsupported OSes with Firefox. I'm currently reviewing the Firefox version history page and porting the remaining EOL info into it, someone looking for those older versions can navigate around the Mozilla archive using that article as a reference. Paianni (talk) 13:36, 8 December 2017 (UTC)
 * As long as you censor information, your layout is unimportant. Wikipedia is not censored.  Fleet  Command ( Speak your mind! ) 09:51, 9 December 2017 (UTC)
 * The practice of using programmatic version numbers to refer to a version of Windows is so stupid that one of Microsoft developers, Raymond Chen, calls the people who do it "poseurs". He says: "Trust me, nobody on the Windows team calls the products by their programmatic version numbers. Whenever anybody says "NT6" I have to go to Wikipedia and look up what they're actually talking about. If I even care to bother, and usually I don't." If you are not writing in Wikipedia for the understanding of others, please don't write at all.  Fleet  Command ( Speak your mind! ) 10:25, 9 December 2017 (UTC)
 * Hi again
 * I just looked at revision 814459896 and I am truely baffled as to what you think and how you think it. But nevertheless, it looks completely meaningless to me. Feel free to print this table and hang it in your room, if you yourself think it makes sense, but in Wikipedia, you are supposed to edit for the understanding of other people not for yourself.
 * Best regards, Codename Lisa (talk) 05:47, 9 December 2017 (UTC)
 * Whatever it is, it makes my screen reader mad. WP:ACCESSIBILITY is being ignored, as well as WP:NOTCENSORED. — Fleet  Command ( Speak your mind! ) 09:52, 9 December 2017 (UTC)
 * Oh, God, this is frustrating. He doesn't get the point. I am staying a day away from this issue to clam down and see what I should do with it.
 * Best regards, Codename Lisa (talk) 06:16, 10 December 2017 (UTC)
 * Well, I am not. He is refusing to get the point and I am not beyond reporting such a person.  Fleet  Command ( Speak your mind! ) 11:54, 10 December 2017 (UTC)
 * My first attempt at the concept was in the, this was the table where it made the most sense as listing the final versions for all the BSD releases and all ISAs was impractical (impossible to a certain extent, as FreeBSD does not host all of the packages for its older branches) and not particularly useful, but placing the emphasis on ISAs was worthwhile as it avoided 'skipping' versions between rows, which would have caused the list to appear incomplete. Three days later it is still uncontested. The following and mobile tables were less successful as unfortunately the old macOS, Windows and Android releases were too prominent to be ignored, but incorporating them would have meant breaking the time order of the latest releases per platform always residing at the top of the respective sections. Now that the concept has been thoroughly rejected, I have come to the realisation that all the tables don't offer enough flexibility for information that has to encompass platforms that are developed very differently, I intend to deprecate all of them and replace them with prose.
 * By the way, one of Codename Lisa's recent edits removed the IA-32 descriptor from a bunch of Windows releases in the older table format. For the record, they are incorrect as while Win9x releases were 16/32 hybrids, they only ran on IA-32 processors and all Firefox builds for Windows until 2015 were IA-32 only, not IA-64 or any of the other ISAs NT has supported in the past. Paianni (talk) 12:59, 10 December 2017 (UTC)
 * I make it simple for you to understand: I oppose removing essential historic data from this table. Everything that you have said so far has been irrelevant, off-topic, technically inaccurate and unjustifiable self-praise. Try addressing the matter at hand for a change.  Fleet  Command ( Speak your mind! ) 13:20, 10 December 2017 (UTC)
 * I addressed the matter earlier on when I moved all the historic data to the Firefox version history article. Despite this, I was willing to compromise and restored the MS Windows info very close to parity with the old format. This was also rejected, and I have decided that neither table formats were satisfactory for different reasons, and proses would allow even more data to be incorporated. Paianni (talk) 13:35, 10 December 2017 (UTC)
 * So, your deletion is contested and the compromise you offer still involves deletion? And your so-called "move" is this: . If I do something like this, my purpose would be to irritate my disputing party. Is that what you want to do here?
 * Problems that you your contribution have caused so far and didn't exist before include:
 * Rendering the template completely useless by removing all historic information from it, which were its soul.
 * Accessibility-phobic summarization
 * Using programmatic version numbers (which even Windows developers don't remember) instead of common names of different versions of Windows
 * Writing said obscure numbers in a way that it is impossible to say what they are without first asking you. How would a layman understand what "52.5.2esr (5.1 and later)" means? Now, I do know what it means, because you said it above, but even I don't know what the hell "2.0.0.20 (4.0 (4.1–4.9))" means!
 * Purposelessly using memory architecture names (and calling them ISAs at the same time)
 * Alright, after all this damage, what good did you your contribution accomplishes? Reducing the size of the table by half an inch?
 *  Fleet  Command ( Speak your mind! ) 13:54, 10 December 2017 (UTC)
 * You are framing your rebuttals as a personal attack, your aim seems to be to make me admit that all the ideas in my table format were wrong (whilst showing your bias towards historic data at any cost), period. I don't see the point as I am letting go of the whole thing, this discussion has outlived its usefulness. Paianni (talk) 14:11, 10 December 2017 (UTC)
 * Fine. Edited them out. Now is there an answer?  Fleet  Command ( Speak your mind! ) 14:32, 10 December 2017 (UTC)
 * Any contributions I made from the 7th December onwards are irrelevant as I'm not pursuing the table format anymore. Paianni (talk) 14:43, 10 December 2017 (UTC)

Post-protection discussion
First thing first: Acknowledging the elephant in the room: User:Papsi is now renamed User:Paianni. Alright, moving on...

How many time must we tell you? You are creating this table for others, not yourself. You must make sure others understand it. And what you did to it lately is as confusing as hell. Why do some operating system have more than one couple of support dates? How is that even possible. Why do you have "all architectures" headers anyway? (That was rhetorical question. You mustn't have. Old macOS don't even run on x64.)

Alright. The page is now protected to encourage you to discuss your changes before committing them. But if you do one more stupid change after the protection is expired, I will personally request that you be permanently banned from changing this template. And one more thing: I am sick and tired of having no one else but the three of us here. I am publicizing this discussion, so that more Wikipedians can review it.

 Fleet  Command ( Speak your mind! ) 07:54, 16 December 2017 (UTC)
 * In one of your Firefox main page revisions, you stated:
 * "Reverted yet another unexplained deletion of the table by User:Paianni"
 * I discussed the proposal in detail with User:Codename Lisa in the talk page from the 11th December onwards, however I was not aware that direct links in the body of the article were discouraged. If the templates have to be retained with all the current info, there are several details I will insist on deploying:
 * The ESR support data should be combined into one row, because there is no need for separate rows for ESR releases where the commencing support dates are identical. In Windows 7 (etc.) and Linux's case, the start of ESR support is different for x64 as a year had already turned when x64 binaries were added to the ESR branch.
 * The 52 (ESR) for IA-32 entry in the Windows section is an unusual case where it represents the end of the line for one set of operating systems and a link in a chain of continuing support for another set. The box should therefore encompass the rows of both OSes, avoiding link duplication.
 * The Latest stable version column should be used to break up operating systems with different preferred binaries (and different support dates) depending on their architecture, in this case 10.5. The previous table had a row titled "10.4–10.5 (PPC), but this is incorrect as readers could assume that 10.4 never had an Intel release. The "All architectures" heading referred to the fact that the binaries listed will run on all architectures supported by a given operating system, e.g as 3.6.28 supported everything that 10.4 could run on, there was no need to label it. In the Windows section most releases were labelled IA-32 but as most entries only dealt with one architecture, there was no point in breaking the column up except for the aforementioned 52 (ESR) situation.
 * A month ago I made a section in this talk page to discuss the arch labels (whether x64 should be x86-64 or not, etc.) but I never had a reply and I deleted it later as I lost interest. How can I count on sections I write being read by others before the changes are made? How can I count on other readers actually comprehending my intentions without visual cues? Could I put draft templates in the Talk page body? By making direct changes to the original templates I am ensuring that other users are bound to see them. Paianni (talk) 12:15, 16 December 2017 (UTC)
 * Disagree with all three! Everything you have written so far arise from the fact that (1) you are already familiar with the release compatibility history (or you think you are) and (2) have little tolerance for extra table cells, and even tables for what matters. These two are in direct conflict with our purpose: to educate people who don't know the release compatibility history. Merged cells can rapidly turn a table into an accessibility nightmare. To top it off, there are your own irritating idiosyncrasies: You write "All architectures" in the header; one might think you mean "architectures mentioned in the table" or "all architectures in existence", but actually, you mean "whatever specified in parentheses". The problem is that in case of Windows 9x and most macOS versions (except 10.5), nothing is specified.
 * Your intention might be noble. But the result of your work has so far not been worth keeping. That's because either you don't think, or otherwise you think about how to negotiate what you like to see, as opposed to what is educational.
 * As for what Codename Lisa told you, please take it up directly to her. I am not here to answer for what she does. (Still, WP:EL is our external linking policy. You might find it useful.) Deleting your own talk page thread was a big mistake. Doing so signals that you have withdrawn any and all objections you had. But of course, you did well retracting any discussion about the use of "x64" vs. "x86-64"; you risk running afoul of ArbCom verdict. See MOS:STABILITY.
 *  Fleet  Command ( Speak your mind! ) 13:24, 16 December 2017 (UTC)
 * I will concede the removal of separate ESR rows was a mistake, my next edit won't involve this. One set of merged cells is enough for this table.
 * For Windows 9x and most Mac versions, there is no need to specify anything; 95 only supported IA-32, therefore Firefox could only support 95 under IA-32, labelling the arch in the row is pointless. All of the Mac binaries supported all the architectures each macOS release supported except for 10.5, where PPC support ended earlier than x86, so the row was broken up. macOS is an odd situation where the 32-bit kernels from 10.5 onwards facilitated running x64 applications, but a 64-bit kernel didn't arrive until 10.7.
 * I used an IA-32 label on the Windows 2.0.0.20 binary because it resides in the row that covers NT 4.0, which supported Alpha for its entire lifecycle, so the arch labelling is required to quash any notions of Alpha, MIPS or PPC support the table might imply for the OS otherwise. The only mistake I made was labelling the versions linked to Win 2000, which was IA-32 only.
 * I was not suggesting that I used User:Codename Lisa to justify my actions, my point was that I had explained my decision in the dialogue there.
 * I had effectively withdrawn my objections to the arch labelling. Paianni (talk) 14:49, 16 December 2017 (UTC)
 * Hello,
 * Your message smells trouble. It demonstrates that you have not understood the problem yet: You don't know what is this table showing and why.
 * Best regards, Codename Lisa (talk) 08:16, 17 December 2017 (UTC)
 * My tables or the predominant table? Paianni (talk) 13:20, 17 December 2017 (UTC)
 * The Fennec and Firefox contrib templates have reverted to their old formats (the latter updated), courtesy of yours truly. I'm sure you or User:FleetCommand will revert them anyway as all my edits are inherently evil, but hey, it's more entertainment for me. Paianni (talk) 16:26, 17 December 2017 (UTC)
 * I deliberately didn't touch those templates. The principle is: When an editor knows for sure that his or her edit is controversial, he or she must not be bold about it. Touching those templates would only hurt your feelings. But as for reverting everything you did here, and then protecting it, this is the nature of Wikipedia: You boldly make mistake and we boldly correct you. One day, someone else will do this to us. (Actually, it has been done to me several times in the past. Once, my colleague aggressively reverted me thrice in a row, with very sound reasoning; I remember being very happy about having such a knowledgeable friend. 2013 was a great year.)
 * The curse of your contributions is that I even know what it is supposed to mean. I know that, for example, the last time, you intended to separate support dates for different hardware platforms per OS. (Although, apparently, you don't realize or don't care how meaningless this is.) I even know which part of the table corresponds to where. But I had no evidence for it and I didn't know you got from the former to the latter. Worse, I have no clue how am I supposed to expand the table in the future if another OS was released. It was not education table, but rather, a cryptography puzzle.
 * An educational table is like a joke: If it takes a lot of efforts to understand it, then it is not a good one.
 * Best regards, Codename Lisa (talk) 20:57, 17 December 2017 (UTC)

Mac OS X 10.6-10.8
I've corrected the maximum supported version of Mac OS X 10.6-10.8. Although Mozilla wasn't too clear on this, the actual maximum supported version was 48.0.2, not 47.0.1. This is corroborated by multiple sources, and I was able to verify this myself on a Macbook running an unmodified version of Mac OS X 10.6.8, with the latest version indeed being 48.0.2, not 47.0.1. Herbfur (talk) 20:24, 7 June 2020 (UTC)

Layout changes since 19th November 2021
This is bitterly ironic for me to say but I prefer the table layout that existed from 2018 to the aforementioned date, before the wikidata integration and further shuffling around. It doesn't help that it's no longer consistent with the mobile and non-official templates either. Paianni (talk) 16:57, 14 February 2022 (UTC)