Linked by Thomas Leonard on Tue 16th Jan 2007 00:32 UTC
General Development In the Free and Open Source communities we are proud of our 'bazaar' model, where anyone can join in by setting up a project and publishing their programs. Users are free to pick and choose whatever software they want... provided they're happy to compile from source, resolve dependencies manually and give up automatic security and feature updates. In this essay, I introduce 'decentralised' installation systems, such as Autopackage and Zero Install, which aim to provide these missing features.
Thread beginning with comment 202532
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[2]: the problem is...
by Obscurus on Wed 17th Jan 2007 02:44 UTC in reply to "RE: the problem is..."
Obscurus
Member since:
2006-04-20

"I've not seen a valid reason for keeping source closed yet, unless maybe it's national-security-related. (And even then, as in the recent spats between the US and UK over their fighter-jet software, some degree of openness may be required). If you can introduce me to one, however..."

Uhm.. it's a pretty big one actually - it is called making a profitable living form the software you write. While some software is well suited to being commercially provided as FOSS on the basis of selling support for the software, in most instances this is not the case. And the model of "we will give you the app + source code for free, and we will sell you support" creates an inherent incentive for the software developer to deliberately create substandard software that requires users to purchase support. Good software should be so elegant and simple to use, so well written and documented that buying support is not necessary.

Not every software developer can afford to spend their spare time writing software, and commercial software companies need a stream of revenue to fund the salaries of programmers. Closing the source code prevents others from:
a)reducing your competitive advantage by using ideas you may have invested a lot of money developing
b)stealing the focus from a project by forking and fragmenting it (as has happened to a lot of OSS projects, particularly Linux).

"On the contrary, it's all wrong; Magical Microsoft OSes are barely usable on a clean install (ok, technically the OS's are barely usable with the computer fully loaded, but you know what I mean)."

No, while I have many gripes with Windows, I am very glad that they provide me with a fairly clean slate to start from - provided I have a web browser, I can add everything I need, and I have little to remove that I don't. I have quite a bit of control over it. I can also slipstream an installation disc so that I can set it up how I like from a single installation.

Now, there are plenty of minimalist Linux distros that also come with a similarly blank slate, but the problem is that I can't easily and painlessly download the few programs I want to use and install them without having to go through any number of time consuming or irritating processes. If you know of a distro which comes pre-installed with Klik or autopackage and a basic gui + browser and little else, let me know.

"...your Windows-worshipping FUD is invalid."
Where did you get the idea that I was a Windows worshipper? Certainly, there are things I like about Windows, but there are just as many that I don't (such as the needless pandering to backwards compatibility, bulky installation size, the GUI etc). Similarly, I very much enjoy using Linux (Xubuntu is my preferred distro at the moment), but there are a number of things that shit me about it (and I have already discussed most of what I feel can be improived in Linux). Same with OSX. Nothing is ever perfect, and openly dicussing the good and bad points of each operating system without fear or favour hardly constitutes FUD.

"Nevertheless, I'm not going to sit here and let you (or anyone else) dictate my choice of app, thankyou."

Where was I dictating what apps you use? Of course, you can use whatever you like, and nothing in what I have said would prevent you from doing this. It sounds like Linux as it is suits you well, and that is fine. I am talking about the one thing that is really holding back linux from widespread adoption (over-reliance on package managers coupled with poor binary compatibility between distros).


"I wasn't aware you were Everyone. I also wasn't aware Wonderful Windows programs didn't spread cute little bits of themselves all over the filesystem (particularly if you attempt to install them on any drive other than C:) and the registry. "


I never said I was, however, the fact that Windows and OSX remain the two most popular OSes has as much to do with Linux's fragmented and unfocussed chaos as it does with dodgy OEM bundling on the part of Microsoft and Apple.

And good Windows programs don't spread little bits of themselves around the PC, many don't even use the registry (most of the apps I use are self contained in their own folder)(granted, there are plenty of badly designed Windows apps that do horrible things to your system). Linux apps tend to spread themselves across a whole bunch of directories (something GoboLinux aims to fix), so I don't think you can honestly claim that Linux (in general) has an advantage over Windows or OSX on this point.

"Translation: Linux needs to become Windows. Except that Windows is not exactly simple (let alone rational), and it's "package management" is even LESS "organised and integrated"."

No, it doesn't need to become Windows, rather it needs to become more focussed and streamlined, and simpler to use by people who have better things to do with their time than fiddling with command lines and .conf files.

Windows'package management (Windows Update) is highly integrated into the OS, only deals with the core Windows components that ship with the OS, and doesn't affect other third party apps. You couldn't make it more organised or integrated.

As I said before there are things about Linux and Windows that I like, and if you combined them into a single OS and discarded all of the bits I don't, I would have an OS I could be very happy with.

I prefer to manage my apps my self, and let the OS take care of itself. Hence my desire for an operating system with an XFCE-like DE that keeps the apps separate from the core OS functionality.

Edited 2007-01-17 02:47

Reply Parent Score: 2

RE[3]: the problem is...
by John Nilsson on Wed 17th Jan 2007 22:37 in reply to "RE[2]: the problem is..."
John Nilsson Member since:
2005-07-06

And the model of "we will give you the app + source code for free, and we will sell you support" creates an inherent incentive for the software developer to deliberately create substandard software that requires users to purchase support

You are probably correct. Concider this:
Usability: "The extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use."

So to design a usable product you have to spcify its users, their goals, and the context of use. The sharper this specification is the better you'll be able to design a usable software.

This means that you have to have quite a narrow focus to create a really usable software.

Now, if your business model is to sell software, the cheapest way to produce it is to copy allready created software. Thus the way to earn money is to create a cheap "orignial" that can be sold to many customers at a high price.

The "many customers"-part isn't exactly compatible with the "specified user, goal and context"-part though. So you compromise and try to creat cheap software for "all users, goals, and contexts". Which in the end means unusable software.

So the "selling copies of proprietary software" isn't really the way to "[g]ood software [that is] so elegant and simple to use, so well written and documented that buying support is not necessary" either.


Now if you base your business model on producing cheap software that is so narrowly focused that no competitor could use the same copies (because you allready saturated the market for that software), the bast way to gain profit is to share a common platform for base functionallity with your competitors and be the best producer of narrowly focused software based on that platform for a select niche of users, goals and contexts.

Reply Parent Score: 2