From: Jeffrey J. K. <bac...@ko...> - 2012-02-10 17:22:40
|
Timothy J Massey wrote at about 11:27:36 -0500 on Friday, February 10, 2012: > "Jeffrey J. Kosowsky" <bac...@ko...> wrote on 02/10/2012 10:33:38 > AM: > > The point is that both 'compress' and BackupPC use 'zlib' compression, > > hence rather than creating some non-standard new suffix, Craig chose > > to use an existing standard to signal that it similarly uses > > zlib. Again, you may have different tastes, but what Craig did makes > > perfect sense to me by reminding me that the files are indeed > compressed. > > Hey, JAR files and ODT files use LZW compression, just like ZIP! Why > don't we use ZIP as an extension for them? It would be so... elegant! > > The point of an extension is to tell the poor, lowly human what type of > data is contained in that file in a glance. This is truly getting tiresome. You have been told by multiple people, multiple times that the "poor, lowly human" is not supposed to be reading these files directly. "Poor, lowly humans" are expected to use the documented GUI. For those who want to delve into the internal file structure of BackupPC, Craig has been kind enough to *signal* compression using the .Z extension. If anybody is too confused by this or does not understand the structure of BackupPC well enough to know how to read such files, then that person should probably stick with the GUI. The developer, and in particular an open source developer, has no obligation to make all the inner workings coincide with your idea of how internal files should be named. When I first started playing with BackupPC it took me all of about 10 seconds to figure out that cat/gzip/compress etc. did not work on these files and that I had to use the *included* BackupPC_zcat utility. It took orders of magnitude less time than this thread flaming Craig for the "dumbest" idea. > If someone creates a random new file format (kind of like > BackupPC's compressed log format!), they should *NOT* recycle a > very well understood extension so that the poor, lowly human won't > be able to figure out why the canonical tool for working with that > format doesn't work (which it does *not*) without some sort of > magic knowledge. The fact that zLib was (ab)used to create this > new format does not mean you should use .Z, any more than > OpenOffice (or Java or any of the dozens of new formats that ZIP > their multi-file contents) should call their files .ZIP. > > And at least in the case of OpenOffice, et. al, the canonical ZIP > management tool *WILL* *ACTUALLY* do something productive, unlike > BackupPC's .Z log files and the canonical tool for managing .Z files. It's an *internal* file -- not meant to be accessed by every day users. Those that want to delve in have never had any problem figuring it out before. It's truly not worth complaining about. Personally, the last thing I want is to have another three letter extension I need to remember. YMMV. If it truly bothers you so much then submit a patch or fork the code and use your own naming convention. There are many more important things to worry about and much better uses of limited and valuable programming time for BackupPC. |