John Siracusa on Delicious Library for the Mac:
As is customary, the background image helpfully includes the “installation instructions” (if you can even call them that). [ those being “drag the icon to your application folder to install”—sil ]
I’m pointing these things out not because Delicious Monster is unique among Mac developers in the quality of their artwork and their attention to detail, but because they aren’t unique. Nearly every popular Mac OS X application is a single-icon drag-installed affair, sporting an attractive icon, distributed in either an internet-enabled or meticulously decorated and arranged disk image. Even open source applications like Fire and multi-platform ports like Mozilla meet this standard on OS X. Heck, even Real gets it right. Real software…think about that!
The is an example of the best kind of peer pressure. There is simply a “climate of excellence” on the Mac platform. Any developer that does not live up to community standards is looked down upon, or even shunned. Commercial, open source, freeware, shareware, it doesn’t matter: pay attention to detail, or else.
Windows users, think about what your typical download and installation experience is like. How many dialogs are you presented with? What do the file names and icons look like? Do you have to run an installer? What kind of manual clean-up is required afterwards?
Linux users, when you look at the carefully laid out disk image contents in the screenshot and links above, think about how far “desktop Linux” has to come before it can even begin to think about details like how single-icon drag-installed applications are arranged in their disk image windows.
Yes, I know, all of this is “pointless” and “dumb” because looks are meaningless. It’s the software that counts—the code, the bits, not the packaging, right? And so we come to an important difference between Mac enthusiasts and other computer users. Mac users understand that the packaging counts too (and are willing to pay for it). Happily, you get a lot of nice things “for free” on the Mac platform these days: composited windows, large icons, rich disk image and application bundle standards, etc.
And why don’t we have this on Linux? Well, the ROX Desktop does, which is why I always liked it so much; it just doesn’t have a lot of the other stuff that makes a good desktop environment, and it doesn’t have enough applications (or enough developers). And we’re not going to get it on Linux either, not with the packaging system in the mess that it is. Now, one big hope for sorting out packaging (if you ignore everyone saying “make Red Hat and SuSE use .debs”—Jono, I’m looking at you here) is autopackage. And what do they say about single-drag-and-drop installation?
What’s wrong with NeXT style appfolders?
One of the more memorable features of NeXT based systems like MacOS X or GNUstep is that applications do not have installers, but are contained within a single “appfolder“, a special type of directory that contains everything the application needs. To install apps, you just drag them into a special Applications folder. To uninstall, drag them to the trash can. This is a beguilingly easy way of managing software, and it’s a common conception that Linux should also adopt this mechanism. I’d like to explain why this isn’t the approach that autopackage takes to software management.
…
Update: As of 21st May 2003 I’d consider AppFolders broke on MacOS as it appears that most Mac software is now shipped using installers, even Apples own software. Wise is also forging a business in InstallShield style wrappers. It seems they really are too simple at this stage in the game, even for the Mac.
Gordon Bennett.
I think you’ve hit the nail on the head why Linux on the desktop is (to me, at least) such a miserable failure so far. (God, am I going to get flamed for saying that).
If you want further proof of how the packaging matters, look at Firefox. A lot of the functionality is the same as when it was still called Phoenix, but now we have an attractive interface theme, consistent iconography and a killer logo. Now it’s really taking off, and I’m sure it makes it a lot easier to get people to have the patience to give it a chance when they see that it looks good as well as works.
Apple, and Mac developers, realise this too. Windows shareware developers all too frequently don’t.
Posted by Ben on November 8th, 2004.
Just to say that the above Ben is not me ;)
/me finds package handling nice and easy in Debian…… but I can see what they’re saying, and the concept is good.
Posted by mrben on November 9th, 2004.
Nope, the above Ben is kapowaz Ben. Plus, he knows how to add http:// in front of a URL, heh heh. :)
Posted by sil on November 9th, 2004.
fsck. Autocomplete from another site is my excuse :(
Posted by mrben on November 9th, 2004.
Just out of curiousity, what’s missing from ROX? I have my own list I’m slowly working through (very… slowly…) but I’d be interested in what you see missing.
Posted by rds on November 12th, 2004.
Missing from ROX: any applications at all that are decently useful and support drag-and-drop saving. Thunderbird doesn’t support it; nor does Mozilla. (Moz does support dragging a URL, sort of, but that doesn’t cut it.) It’s all well and good having Edit do stuff, but it’s a toy editor. Yes, there’s rdsEd, but that really came along after I got out of Rox :)
An IM client. OpenOffice support for drag-and-drop saving. A decent amount of active developers who are doing the sorts of things that the Gnome project are doing, like Beagle and Dashboard and udev/hal/gnome-volume-manager. Proper, real applications packaged as AppDirs: not AppDirs that link to your previously installed packages, but real AppDirs. Some kind of solution to how we handle dependencies in AppDirs. A revised filesystem layout, a la GoboLinux. Really, I think that the ROX Desktop is too thin a shell to go on top of a standard Linux distro; there are too many holes in it where you look through and see the underlying OS, and it’s totally un-ROXy. What’s needed is the ROXLinux project to work, with Applications and Library directories and LibDirs and the like, and it never really got off the ground.
Posted by sil on November 12th, 2004.
Sounds about right. (Though I think you give rdsEd way to much credit. ;) ) Though I think a good number of those gripes could have been avoided if they would just have started with the idea of slowly moving in one idea at a time. Say, start with a Filer that did AppDirs, get everyone saying “hey, that’s nice,” and get some AppDirs out there. Then slowly move in, say, DnD saving.
Posted by rds on November 13th, 2004.