Le 18.10.2013 04:32, Miles Fidelman a écrit :
berenger.morel@neutralite.org wrote:So, what you name an OS is only drivers+kernel? If so, then ok. But some people consider that it includes various other tools which does not require hardware accesses. I spoke about graphical applications, and you disagree. Matter of opinion, or maybe I did not used the good ones, I do not know. So, what about dpkg in debian? Is it a part of the OS? Is not it a ring 3 program? As for tar or shell?Boy do you like to raise issues that go into semantic grey areas :-)Not specially, but, to say that C has been made to build OSes only, you then have to determine what is an OS to make the previous statement useful. For that, I simply searched 3 different sources on the web, and all of them said that simple applications are part of the OS. Applications like file browsers and terminal emulators. Without using the same words for the same concepts, we can never understand the other :)I'm pretty sure that C was NOT written to build operating systems - though it's been used for that (notably Unix).
I never said I agreed that C was designed to build OS only. Having to manage memory does not make, in my opinion, a language made to write OSes. I never took a look, but are linux or unix fully written in C? No piece of asm? I would not bet too much on that.
It was simply to argue that, even if it was true, then it does not avoid it to be good to write applications, since, in more than one people's opinion, OSes includes applications.
I agree, it is part of programmer's job. But building a bad SQL request is easy, and it can make an application unusable in real conditions when it worked fine while programming and testing.Sure, but writing bad code is pretty easy too :-)
I actually have less problems to write code that I can maintain when I do not have to use SQL queries and stored procedures. I simply hate their syntax. Of course, I am able to write some simple code for sql-based DBs, but building and maintaining complex ones is not as easy as building a complex C++ module. Or as an asm or C, or even perl. Of course, it is only my own opinion.
I'd simply make the observation that most SQL queries are generated on the fly, by code
It is not what I have seen, but I do not have enough experience to consider what I have seen as the truth. But I will anyway allow myself a comment here, if some people wants to create a new language: please, please, please, do not do something like powerbuilder again! This crap can mix pb and sql code in the same source file and recognize their syntax. It will work. But will be horrible to maintain, so please, do not do that! (this language have really hurt me... and not only because of that)
But now, are most programmers paid by societies with hundreds of programmers?(and whether you actually mean "developer" vs. "programmer")I do not see the difference between those words. Could you give me the nuances please? I still have a lot to learn to understand English for precise terms.The terminology is pretty imprecise to begin with, and probably varies by country and industry. The general pecking order, as I've experienced it is (particularly in the US military and government systems environments, as well as the telecom. industry): Systems Engineer: Essentially chief engineer for a project. (Alternate term: Systems Architect) - responsible for the big picture - translate from requirements to concept-of-operations and systems architecture - hardware/software tradeoffs and other major technical decisions Hardware Engineers (typically Electrical Engineers): Design and build hardware (including computers, but also comms. networks, interfaces, etc.) Software Engineers: Engineers responsible for designing and buildingsoftware. Expected to have a serious engineering degree (sometimes anEE, often a Computer Science or Computer Engineering degree) and experience. Expected to solve hard problems, design algorithms, and so forth. Also have specific training in the discipline of software engineering. People who's resume says "software engineer" have typically worked on a range of applications - the discipline is about problem solving with computers. Developer: Vague term, generally applied to people who do the same kinds of work as software engineers. But doesn't carry the connotation of an EE or CS degree. Tends to be commonly used in product development environments. In my experience, a lot of developers start out writing code in their own field of expertise (doctors writing medical applications, musicians writing music software, and so forth). People with "developer" on their resume often have specialties associated with the term - e.g., "game developer" - emphasizing an area of specialization. Programmer: A term I've never actually understood. Basic use seems to be "someone who knows how to program" or "someone who programs for living." But.... I've personally never seen anyone hired purely as a "programmer." (I've hired, and seen hired, a lot of developers and software engineers, but never a pure programmer.) As far as I can tell, the term is a carryover from the early days of large systems - where "systems analysts" figure out what a system was supposed to do (i.e., did the engineering) and then directed programmers to write code. (Akin to the relationship between a hardware engineer giving a design to a technician, who would then build a prototype.) Coder (or "code monkey"): A somewhat pejorative term for an unskilledprogrammer - someone who might have taken a couple of "introduction toprogramming" courses and now thinks he/she can write code. For what it's worth - my observation is that the demand is for software engineers and developers (i.e., skilled people who can solve real problems). But... computer science curricula, at least in a lotof schools, seems to be dumbing down -- a lot of places look more likeprogramming trade schools than serious engineering programs. Not a good thing at all.
It seem you attach a lot of importance to schools ( that what I understand in "computer science curricula" ) but I wonder: how do you consider people who learned themselves, before even learning a craft at school ( that's what wikipedia says when I start from the French "métier" and then switch on English language. I understand that it seems to be reserved for physical products in English, but, I like that notion: "A craft is a pastime or a profession that requires some particular kind of skilled work." which makes things different from a simple job. ) ?
Would you name them coder, programmer, developer? Something else? Do you consider they are always less efficient that people with high degrees?
And, again, just a guess, but I'm guessing the huge percentage of programmers these days are writing .NET code on vanilla Windowsmachines (not that I like it, but it does seem to be a fact of life). A lot of people also seem to be writing stored SQL procedures to runon MS SQL.Windows actually provides a stable environment, very good programming tools ( visual studio is a really good IDE ), and when you write for it, you know that most computers will be able to run your program.Well sure - but, by definition, it's not cross platform, and I sure wouldn't write server-side stuff to run on Windows (though many do, including my current major client).
If one day, I can work on something else than on windows, I could accept lot of sacrifices for that. And I am probably much more efficient outside of windows world, or to be exact, in writing portable stuff, than on writing stuff for one OS only. But, the truth is, while MS Windows will be running on more than 70% of computers, we will have to write and maintain for it and it will be used for servers, too, because it is easier to do so when there are less targets. Same for applications written in multiple languages: the less languages you use in a software, the easier it will be to maintain, and so, the cheaper. On the other hand, portable softwares forces to have better design, because you really need to separate the layers, which in turns helps to find and fix bugs faster, imo.
I probably wouldn't write a telecom. switch control program to run under Windows, for example - and I expect visual studio wouldn't help a lot in writing Erlang code (though I'm not sure there's a good IDE for Erlang).
In my last job, most of computers and servers were running windows. I think there were some centOS servers, but unfortunately the internal softwares were written in powerbuilder, with work to rewrite them in C#/.NET. I can remember the pains of making powerbuilder able to access a webcam ( through C++ and a poorly documented API ). If you think there is no good IDE for Erlang, then, try to use PowerBuilder's own IDE to write PowerBuilder code. You will change your mind, and even the simplest text editors (yes, including the ones without syntax coloration) with a simple file browser will seems like paradise then. At least, those ones will be stable.
Arguably, it is also easy to install stuff on windows.I strongly disagree with that for most current linux-based OS, but I remember my first installs of Debian. I was used to consoles, but I had no knowledge of what tools I had to do anything, and no graphical interface to learn basics from the Internet. Debian without Internet was not usable for beginners at that time (installer provided woody/potato at that time IIRC. I had my first successful use of Debian with Etch... or was it Lenny? I only know that I know all those version names for a reason or another, don't remember completely). So, that linux-based reputation is not stolen, and since it is not used a lot, it has not been reduced unlike the windows' one to be unstable and full of security holes. And when people asks me what I am running with, I have to reply it is a Debian, and insist on the fact I have heavily customized it to avoid frightening them.Those are probably the reasons why people programs with Microsoft's "closed" technologies. But now, MS can argue that .NET is available on linux too, and that SQL is standard ( but not procedures which are an important part, of course ). We can just hope that windows usage will continue to very, very slowly regress to have more portable softwares.Are you saying it's NOT easy to install stuff on Windows?
Compared to other things I know, aka, deb packaging system, yes.Considering you already have the archive binary, on windows you will have to read lot of screens, or you will destroy your system in less than 1 year, if you often add and removes applications, because when you remove something which installed crap on your computer, it won't remove it.
Sometimes, on Debian, I have to compile stuff myself... but it is the same for windows.
And, finally, to install and configure programming related stuff, like development libraries, on Debian you simply install them. They are installed in correct places so it is automatically detected by tools which needs them. On Windows, when you will have finally installed it, you will have to take care where you did, because you will need that information later, and all softwares have their own preferences. Not to speak about the fact that they usually install in "c:\program files\<company name>\<software name>\". The company name is sometimes editor's one, or developer's one. Random.
Installing stuff on windows is only easy when you only know windows in modern OSes (this is to discard ms-dos ;) ).
Oh, and I did not mention that, on windows, installing a software which depends on another software will need you to install manually that other one. Which is why softwares often (but not always, I had lot of errors claiming that something was missing) includes most (but not always all) dependencies.
So, yes, generally, installing stuff on Windows is hard, compared to the only other modern system I know: Debian. Oh, and, I do not use any DE ( even file browsers lacks on my computers, they are too slow against terminal emulators ), so, yes, I need the console to install stuff. I can not be considered as a normal end-user, it's obvious. But, at least for Ubuntu, there are means to install a software directly by clicking a link on your favorite web browser (I noticed that on few websites). I guess that this technique uses xdg-open in background, and if I am not wrong, it would mean that file browsers are also able to do so. As for exe and msi on windows.