I don't think synaptic can handle RPMs. Plus, its UI sucks. The only reason people still use it is familiarity. If synaptic was released in this state today, its developers would be shot.
Synaptic's UI is fast and intuitive. If you're a seasoned or semi-intelligent user of .deb-based distros (and a bad typist), it makes most package management tasks quicker than typing. I haven't seen a good RPM GUI yet (though I haven't played with mandrake/mageia). Yast is painfully slow nowadays and packagekit is.. ungood.
One of the things that has always bothered me with Software centers in Linux has always been that there is no single software center, nor is there a universal package manager. Something I would love to see done (I don't yet have the skills for it, so I can't work on it quite yet) is to have a software center that isn't controlled by any one company or community, that would (or at least would be great if it was) taken up by most distributions, and have their package managers tie into it in some way so that you could interface with it almost any way you wanted to but still have access to the packages everyone else had. That way people could choose their own package managers without worrying about not being able to get a piece of software under a new manager that they could with their old. Another good thing about that would be that their wouldn't be different "worlds" under each distribution where they keep different versions of different software. I dunno, just throwing my two (or maybe three) cents in.
you DO realise that what you're proposing is to destroy all GNU/Linux distribution while creating One True GNU/Linux (hint: packages for different distribution are incompatible because they not only binary incompatible/built differently but also, well, _packaged_ differently which means that they have different sets of files put in them) ?
also, praising interpreted languages over natively run ones is bullshit, people. just admit that you're lazy fucks and want some "magical" mumbo jumbo to take care over things you don't want to spare your brain cycle on, cutting corners and making shortcuts in your work.
You can write fast applications in Python and Ruby.
It's definitely possible to write a software center app in one of those languages that starts up quicker than a C++ application like Firefox.
The reason why Java and C# apps all start slow is because they have to load their huge frameworks. And if the user doesn't have them installed yet, install 100+ mb of crap (number for java or c#). How does this not apply to other interpreted systems?
Originally Posted by BlackStar
Those evil VMs/interpreters that are stealing our precious CPU cycles, slowing down the very CPU intensive software-center GUI. Damn them! Not to mention other stuff.
No, in fact I'd like to know what you meant by "other stuff". What other stuff do those evil VMs/interpreters do, other than "steal your CPU cyclez"? (For some reason I don't think you'll answer "improve security", but feel free to prove me wrong).
Random pauses for GC or other reasons. Users really love it when a button press randomly takes 2 secons to register.
Also taking ram from "real apps", meaning if a workload uses a lot of ram, desktop grinds to halt as it has to page in and out the interpreters and frameworks for n+1 languages.
Windows doesn't have one and it seems to be doing alright...
However I'm in 100% agreeance that small OS's (ok anything that's not Windows) need to do everything they can to interest users and keep them excited. The amount of software available on Fedora/Ubuntu is staggering but I think most (especially new but also experienced) users have no idea. Some sort of "what's new", "what's hot" type listings would go a long way to building that connection between developer/packager and the users.
Except it's the most trashed OS ever with dozens of duplicate crap and terrible installers.