What's the point of Zip files?
rant tech web · 4 comments · 250 words · read ~353 times.
My laptop ran out of space yesterday. Why? Useless ZIP files!
I needed to download a Windows Virtual Machine in order to upgrade the firmware on a device (long story). The official Windows 10 VM is 20GB
TWENTY GIGA-FUCKING-BYTES!!!
It downloaded reasonably quickly - yay fibre! But I had to wait almost as long to unzip the bloody thing. Whereupon, I discovered that zipping the file - and it was only one single file in there - saved a whole 200MB. Yup, a 1% saving.
As it happens, I downloaded the wrong VM. So I downloaded the right one. A similar size, although this had couple of files in it - but the ZIP didn't save much space.
At which point, my laptop - not unreasonably - threw a wobbly because I'd suddenly consumed 80GB of space!
I'm sure you're about to tell me that there's an esoteric Linux command that will automagically extract a file, delete the original archive, and repartition my SSD for optimal layout - but that's not the point.
Every web server can - and should - gzip files on transmission. Manually zipping a single file on your server doesn't save any download time. It doesn't save any bandwidth.
Perhaps it makes sense to bundle a few related files together - but if it is a single large file, a .zip just wastes the time and disk space of anyone who downloads it.
Jude Gibbons says:
Hazarding a guess, the zip files are automatically created so if you are sending a bundle of 30 files it makes sense, but if you are sending only one (and it only saves 1% of the size) you have no human to intervene and skip that bit of the process. I've also had email systems that mangle files (even pdfs) unless they are zipped. So their thinking is: better safe than sorry, and by the time you've downloaded it it's no longer their problem!
@Edent reminds me of "I want to download these 50 JPEGs plz — oh sure lemme zip them up fer ye..." Nooooo
I've seen zip files being used when the file extension of the file is considered suspicious. I know that it would be simple for them to rename it and send it that way, I suspect that the majority of users haven't even turned "hide extension for known files" off. Where as double clicking on a zip file is a common thing for users to do. I'm not impressed with zip, as the other day I downloaded a development framework in a zip, moved it into position and then wondered why my compiler was still saying it couldn't find the headers, it took me a while but I eventually discovered that none of the symbolic links were symbolic links, they were a file with the text of the location.
Tim says:
The .zip format includes 4 bytes for CRC, so that does add a little bit of protection against in-transit data corruption that wouldn't otherwise be there if you just downloaded the raw files.
If downloading the file via FTPS/HTTPS, however, I doubt that CRC adds anything, as the encryption algorithm ensures data integrity.