skeletonbow: If you want to save a lot of space, try using the xz or lzma archive formats, supported by 7zip et al. Doing the compression on a machine with lots of RAM can yield better compression results as the encoder is capable of better compression when given unbounded gobs of RAM. Just an idea to ponder.
rtcvb32: But it's apparent interest in taking advantage of this is rather low priority.
Another highly contributing factor is that for example I own approximately 1/3 of the GOG current game catalogue (242 out of 700 games), and that consumes approximately 340GB as-downloaded (not installed), including all bonus materials for each game. With hard disks going for around $60 or less for 1TB, and cheaper per TB as you increase the size to 2/3TB or possibly larger, the cost per TB is super cheap and there is a strong argument to be had for buying a new (or used) hard disk large enough to just store all the games outright than to spend too much effort trying to compact them for very small benefit of 1-3%.
If someone at GOG can toggle a flag in a compression tool or even use a similar tool without much effort at all and without risk to quality assurance or other problems and gain a non-negligible amount of reduction in size compared to the going price per terabyte of disk space, then hey - why not! But I wouldn't want to see them spend too many man hours of effort doing such a thing for ever diminishing benefit as hard disks and other data storage devices increase at sizes far more exponentially than old games increase in size for example. ;o)
I suspect this is a rather niche thing in that the number of people who could benefit from it is likely relatively small percentagewise of the customer base, unmeasureable without some form of data gathering tool or hardware survey like Steam uses, and of those customers that could benefit from it on old hardware with small storage space only a fraction of them are even affected and only a fraction of them actually care and only a fraction of them who care would end up benefitting.
That's not suggesting to do or not do anything, just trying to rationalize that even if it is beneficial to some subset of users who would actually benefit and care, the man hours to do it out of the man hours of finite resources available are possibly better spent bringing new games to the catalogue, or fixing bugs or something and low impact changes like this could get thrown in a tickler file as low priority for a future time. Kind of like how they just updated all the installers in November, which was a change that was lingering low priority forever but eventually happened. ;o)
I used to experiment very heavily with various compressors and algorithms in the past, ZIP, ARJ, LHA, RAR, UC2 and many others both in DOS, Linux and Windows over time. Eventually I settled on RAR for years and then eventually on zip in Windows and bz2 in Linux even though there were others out there that gave better results for particular types of data etc. The reason was largely that improvements in compression technology were largely being dwarfed significantly by the doubling in size of hard disks etc. every 18 months. Compression wasn't halving in size every 18 months for example, so you end up with a law of diminishing returns and the time and effort put into it was mostly academic than much real world benefit for most intents and purposes.
Now that 7zip is very popular in Windows and available everywhere I might need it, I use that with the highest settings by default, and if they're files to pass to others they always have 7zip already anyway or I can send them the URL. That works for personal usage anyway (although not necessarily for product offerings from a business per se.) And in Linux I mostly use bzip2 still but transitioning to XZ for larger scope things in which the savings are actually beneficial to me or others (ie: creating rpm packages, zipping huge archives of files that benefit from compression in the multi-gigabyte range etc.)
Anyhow, it is still interesting stuff and worth experimenting with over time whether for academic reasons or real world benefits. If nobody ever does it, we might be missing out on something good! :) XZ is definitely the hugest breakthrough in general purpose compression I've seen since Bzip2 though, and where my personal experiments would begin if I were to get the spark. :)
