User talk:TheAntiDisruptor

TheAntiDisruptor, you are invited to the Teahouse!

Hi TheAntiDisruptor! Thanks for contributing to Wikipedia.
Be our guest at the Teahouse! The Teahouse is a friendly space where new editors can ask questions about contributing to Wikipedia and get help from experienced editors like Rosiestep (talk).

Visit the Teahouse

We hope to see you there!

Delivered by HostBot on behalf of the Teahouse hosts

16:03, 6 July 2020 (UTC)

My recent edit to tar (computing)

Good day to You!

You undid my recent edit to the page Tar (computing). I actually encountered the .tbJ file extension when our lab purchased Gaussian 16 – it came in archives of this format. WebMO webpage is just pasting the official Gaussian instructions, which are below:

   If a file with .tbJ extension is found do the following:

   % tar xvfJ $mntpnt/tar/*.tbJ
   % chgrp -R <grp> g16               # <grp> is the group from step 1.
   % cd g16
   % bsd/install

   Otherwise do the following:

   % bzip2 -d -c $mntpnt/tar/*.tbz | tar xvf -

I do not know how to cite the (publically unavailable) README properly, so I just searched the Internet and found the description on WebMO page.

I do not know how significant all this is (maybe it is just some weird internal convention of Gaussian, Inc.), so the decision is still up to You. But I wanted to clarify myself nevertheless.

If You are wondering, they are using csh by default for installation.--Esmu Igors (talk) 09:39, 16 July 2020 (UTC)[reply]

@Esmu Igors: To be clear, changing the filename extension does not change the file format. The entire filename, including the extension, is just metadata. Changing it doesn't change the contents of the file. E.g. changing the filename extension from .jpeg to .gif, doesn't magically convert a JPEG file to a GIF file, you just get a misnamed JPEG file. The standard file command is useful to find out what the actual format is. When it comes to xz compressed tar files (which this is about), .tar.xz is obviously correct and unambiguous. The short form .txz is obvious given .tgz (.tar.gz), .tzst (.tar.zst), et cetera.

This .tbJ business definitely seems to be a weird internal convention of a company that doesn't know what it's doing in this regard. Using .tbJ as a short form for .tar.xz makes no sense. They seemingly got the capital J from the GNU tar option for xz, and probably didn't understand that the b comes from bzip2 (.tbz is short for .tar.bz2). They don't even seem to know about the j option for bzip2, since they manually decompress bzip2, but not xz. Hence why I questioned using them as a reliable source in this matter.

What purpose do these new short forms even serve? It's not like there are many production systems with only support for 8.3 filenames that are routinely decompressing xz files (which can have a 1.5 GiB dictionary). TheAntiDisruptor (talk) 16:18, 16 July 2020 (UTC)[reply]

To be clear, changing the filename extension does not change the file format.

I understand that :) The table had the list of commonly used file extensions, though, so I thought I can contribute.

I do not think, however, that Gaussian, Inc. do not understand something because this is a software company running since 1986 and targeted primarily on Unix servers (which are overwhelmingly dominating the market for computational chemistry). I think the "not using j option" thing is here just because the README is mainly copied and pasted from that of previous versions with necessary modifications (I am glad that for this release they at least explicitly assumed the possibility of system mounting the DVD automatically). As for the b letter — maybe they assumed "tb" is a contraction for "tarball"? Anyway, personally I think there is a certain incentive for mentioning .tbJ because Gaussian is clearly and by far dominating the computational chemistry software market.

It's not like there are many production systems with only support for 8.3 filenames.

Many researchers use programs such as Midnight Commander, so they will typically want names as short as possible (otherwise You get Your good old tildes).--Esmu Igors (talk) 17:53, 16 July 2020 (UTC)[reply]

@Esmu Igors: Is there a reliable source that supports the claim that .tbJ is a "commonly used file extension" for .tar.xz files? Use by one company for distribution of one piece of software in a very niche market is about as far from common usage as one can get. When I searched for ".tbj tar" using Google, the only relevant result was the "WebMO" site about how to install "Gaussian 16". Contrast that with how common .txz appears to be. When I searched for simply "txz" using Google, the entire first page is about .tar.xz files.

Another contrasting example is in software support. If you check out GNU tar's source code (specifically tar-1.32/src/suffix.c), you'll see support for .txz but not for .tbJ. According to the changelog, support for .txz was added on 2013-09-23. I was curious, so I tested it out, and auto-compress does indeed work for .txz files.

The short form .txz is obviously noteworthy, while .tbJ is not. The Gaussian software itself is obviously deemed notable, as it has its own article. That doesn't mean that everything Gaussian, Inc. does is noteworthy. For all we know, .tbJ could be a copy-and-paste accident or typo (e.g. if they thought they edited the tar options and instead unknowingly edited the filename extension). TheAntiDisruptor (talk) 20:51, 16 July 2020 (UTC)[reply]

OK, You are right. Then the discussion is closed. Or maybe we should transplant it to the article talk page?--Esmu Igors (talk) 06:21, 17 July 2020 (UTC)[reply]