суббота, 6 октября 2007 г.

bludgasm: Could it really happen?

I'm hearing more and more about novice users running linux lately, and it really does make me feel good. I truly believe that Ubuntu, not this year, probably not next year either, but by 2009 could really start to push Microsoft out of the way. We'll never be rid of proprietary software, nor should we be. Both proprietary software and free software packages have done an equal amount of innovating and inspiring each other. What I've seen is that free software shows us what users and developers want to have, and proprietary software shows us what enterprises are willing to pay for based on immediate needs and existing education. They are two completely different worlds that work in completely different ways, and yet they shape each other; they need each other. If you discover a market niche you deserve to be compensated handsomely. You've created a valuable piece of of a businesses infrastructure. But that value depreciates over time. It's absurd to have a truly residual profit situation. And eventually you will lose sight of your goals, and you will be stuck on an idea doomed to deprecation. More likely than that, you will have acquired other projects. That's when the open source community takes over your role and designs a fresh work based on the collective needs of everyone. Having a whole world of programmers constantly updating code, auditing security issues, and so on. It happened to UNIX implementations, it happened to web browsers, it happened to media players, it happened to web servers (very quickly), and now it's finally happening to the desktop. With my Arch-Nemesis, DELL and their partnership with Ubuntu, I can see it happening the near future. This is not to say that Free software developers don't innovate. They just don't have the capital (in money or time) to innovate the way that the big players with the heavy market shares do. But I think that is subject to change, once free and open source software begins to dominate the market. At that time, companies will be able to worry less about massive licensing technicalities and can actually hire programmers give them what they need and potentially merge their modifications back upstream. This is when you have created a demand for open-source software to develop unheard-of techniques and masterpieces, and thanks to those who need it rather than those who wish to sell it, it will be done properly. All I can hope for now is for Darwinism to take over the role of what was totalitarianism. May the fittest solutions succeed to de-facto standards quickly and with endurance. We need less fragmented modes of centralized authentication systems. We need to abolish the need web applications that require "ActiveX" without jeopardizing the clients' security. We need a fully-fledged and efficient anti virus system that has a large community as an authority for its virus-definitions repository and it must be libre gratis, because Linux is already known for having an extremely low virus risk, and that should never change. We need more GUI configuration tools and wizards, but NEVER EVER EVER take away my beloved configuration files, because we, the admins, deserve to know exactly what our applications are doing, and should have the right to do out of the ordinary things, or memorize/reference useful configurations and not have to go through several layers of menus to get a standard system up and running. It's just not practical to take the Microsoft approach and attempt to put every possibly setting in a gui and save it in some insane binary formatted XML or in some crazy registry equivalent. You will forget something, and additional features will become unmaintainable. I'm sure there are other things I can gripe about when it comes to the fragmentation of GNU or otherwise free software systems. But I'm not worried about it, because every single year I watch integration get tighter, and tighter and more seamless, and as the few major players of window managers get more advanced and more refined, I see the fragmentation problems practically disappearing with every new version. One thing that I do want to see less of is interpreted or JIT programming. As a technology minimalist and a somewhat of a purist, I think languages like perl and python should be used for custom applications and quick fixes, and things like web applications that need to be updated frequently. But for desktop applications, let's stay away from python In example, not many things are more frustrating than using Nicotine, and that's because it has excellent features and a great implementation, but given that it is interpreted by python, it is a major drawback in performance, especially when managing or browsing very large media libraries, or doing multiple searches, or having a large que for downloads and uploads. There's no reason why this couldn't be effectively implemented in c++ with heavy use of its own libraries, standard libraries and a standard GUI toolkit library. All of the above text in this post is fairly unrelated to what I wanted to write about. Since I bought my laptop over a year ago, I've had a lot of annoying driver issues. By the time debian Etch was released most of them were resolved, but I always had an extremely difficult time getting GL stuff working with my ATI card. I love Debian. I love the structure of Debian. I will only deploy debian on servers. But not even released yet, still in beta, the new version of Ubuntu configured my desktop perfectly. Free Software Purists can criticize all they want, but Ubuntu's modest effort for non-free software really makes a difference when there are no better (free) alternatives. I finally have all the features of compiz and xgl working in harmony. Why should I have to compile my own drivers and configure my X Server when other people with the same hardware have already done so? Willingness to work with what is available (and functional) makes the implementation efficient, and helps us learn and prepare for creating truly free packages that are fully-featured and can replace the commercial packages. I can't wait for the final release of Ubuntu "gutsy". I'd go as far as to say that I can't wait until next April to see what the NEXT version of Ubuntu will be like. I've been running various distributions of Linux since 1998, and I have never seen anything develop and prosper (not to mention just work) as quickly as Ubuntu. Although it's not 100% free software, I really believe that Ubuntu is going to pave the way for completely free software systems.

1 комментарий:

Paddy3118 комментирует...

I don't agree with you about Dynamic languages like Python.

I think that you would find that some utilities are only written because writing them in a Dynamic language is easier. Yes, they may need optimising for performance, but ban Dynamic languages and the range of utilities available to you will shrink.

Remember also that Dynamic does not necessarily mean slow. The algorithm has a large impact on speed and languages like Python may allow you to explore different algorithms for a problem in less time than it would take to implement the first algorithm you thought of in C++.

- Paddy.