JP: No contradiction at all. First, I said sysadmins don't like to have too much work. Second, I'm not really a sysadmin, I'm a programmer. I like to configure it myself, yes, but I also want it to be well configured and not having to change anything until I really need a major update. Do it once, enjoy it many style of living.
10: So, what is then the real problem with Linux? What is your real opinion, is Linux ready for the desktop or not?
JP: As I said, the problem is not with Linux, but with *nix. Linux is not the major problem. In fact, I do believe Linux is ready for the desktop. It's the *nix world that is not, and as such, Linux suffers from that.
11. And what about a programmers point of view, what's wrong and right?
JP: In Windows programming, there are a few things wrong also. For example, there are a lot of functions to do the same thing. And not much API documentation, although it appears to be well documented. We have a huge list of constants and types and a strange naming convention to follow... not really mandatory. And for some simple tasks, there are functions that go through multiple layers before executing the task. In the other hand, there is, let's call it, a strong empahsis on object oriented programming. This is very good. Object oriented programming is excelent. And there are lot's of development tools for visual oriented programming, which is also good, and those tools, usually have good documentation for the API they support/provide. In Linux, it's the opposite. We have a simple and well documented API and enjoy more flexibility. The redundancy is, IMHO, residual. But, it is not an object oriented operating system. The use of C++ and other OOP languages first go thro! ugh C callbacks. That's really not a limitation imposed by Linux, but rather by the architecture it is supposed to support. But in fact, that doesn't cause that much overhead. I would rather program for Linux than Windows, in most of the situations, especially when developing console programs and services. But the GUI development is faster and easier in Windows than in Linux. There are not much visual development tools for Linux, ie, for X, and they are not as good as the Windows ones. The big exceptions are the Borland tools, my favourites under Windows (Borland C++ Builder). Kylix is also excelent, especially Kylix 3 which added C++ to RAD development with Borland tools under Linux, but I still prefer using Glade, GCC (g++), Glademm and GTK. The biggest problem of all, is the big costs associated with developing using such commercial tools. These tools are expensive, and I mean really expensive. There is always Glade and Lazarus.
12. In your opinion, what are the most important technologies of today?
JP: I immediatly remember XML. XML and XSLT. Those are revolutionary. XML brought the solution to an enormous amount of problems, some of the with several decades of existence. Especially in document and content management areas. Also I believe OOP is extremely important and databases also. I could mention the Internet and all the technologies around it, but it is already something of our lives, doesn't have that taste of new thing anymore. That doesn't mean the Internet is not important, because it is. Probably the most important concept right now. PHP is also very interesting.
13. How do you see the technological future?
JP: I see a networked, object oriented, XML database driven and XSL transformed world.
14. How about the hardware? In the future, I mean.
JP: That is a complicated question. There are too much hardware types that can improve, I can see object orientation in hardware too. No like the software, but object oriented in a hardware way. That's one of the projects we have been developing secretly. In the Microprocessors, for example, the CPU developers will be less concerned about having the chip working at higher clocks, and more concerned about having it executing more instructions per clock, decreasing clock rates to decrease energy waste. Also, the CPU development is reaching a point of stagnation, where some architectures that are standards today will suddenly cease to exist, and be replaced by cleaner, improved, simpler, new ones. Hard drives will keep increasing size, SATA will take place in a more or less near future, and some time, not in the very near future, disk harddrives may well be replaced by big flash drives or equivalent technology. The user interfaces at some point will have to evolve. On keyboa! rds, I can imagine they will remain pretty much the same as today, some minor differences. On mice they can continue like today, remaining only the laser mice, or they can evolve to the currently existing touchscreens or some device that can follow eye movement or a finger movement. I don not forsee much future for small wireless devices like mice and keyboards, expecially for the radio based ones, because they are expensive, they drain more power, and they emit higher levels of radiation. I do think wired and wireless devices like these will continue to co-exist, but the wireless devices will continue expensive and restrained to some specific and special uses.
15. And what about the software?
JP: Besides the victory of Object Orientation and XML, I can also imagine software developers understanding finally that any program that does not take full advantage of the resources it has available it is a bad program. That any program that is not speed optimized is also a bad program. And that any program that wastes resources and that takes more space than it should, is not even worth of being called a program. Hard drives and memory will keep increasing in size, and CPU speed will increase also, but not to be wasted with unoptimized programs. Rather, they will be improved to accomodate more user data, and to process it faster.
- "Joćo Paredes Interview, Part I"
- "Joćo Paredes Interview, Part II"
- "Joćo Paredes Interview, Part III"
- "Joćo Paredes Interview, Part IV"
