Talk:Computer software
|
|
Software varieties
I think, this paragraph sould to be like this. Classical division of software in computer science is system and application software. I don't understand devision to platform, user software in top level of classification. Conan 21:05, 7 Nov 2004 (UTC)
In computer science all software divided to two big classes:
- System software
- System software includes the basic input-output system (often described as firmware rather than software), device drivers, an operating system, and typically a graphical user interface which, in total, allow a user to interact with the computer and its peripherals (associated equipment). System software often comes bundled with the computer, and users may not realize that it exists. Programming tools also are part of system software, because compilers, and other tools are always dependent of the operating system and CPU, and all system software can't be build without compilers and linkers.
- Application software
- Application software allows a user to accomplish one or more specific tasks, like office suites and video games. Application software is often purchased separately. Applications are what most people think of when they think of software. Sometimes applications are shipped with a computer, but that does not change the fact that they are applications.User software tailors application software to meet the users specific needs. User software include spreadsheet templates, word processor macros, scientific simulations, graphics and animation scripts. Even email filters are a kind of user software. Users create this software themselves and often overlook how important it is.
See the rest of this discussion page for explanation and references. As far as I can tell, there cannot be any dispute over the fact that there are *many* kinds of software. The real debate is about what the "computer software" page should say. If this page is about "computer software" then it should openly discuss all kinds of computer software, rather than limiting it to jsut the CS point of view. At least that is what I believe. If you want to create a separate page for the narrow-minded CS concept of software, that would would be fine by me. -- The phantom avenger for software engineering
Since this article has such a general title, perhaps it should focus more on the difference between software and hardware than the various different ways of classifying software.
Software classification
I very surprised to see more top SW classes, except system and application. Who can explain me nature of others classes? Kenny 12:52, 2004 Oct 25 (UTC)
- See the rest of this discussion page for explanation and references.
- There is a huge amount of info about end user programming. Try link 1 (http://www.stsc.hill.af.mil/crosstalk/2004/06/0406Cook.html) or link 2 (http://www.computer.org/software/homepage/2004/July-Aug/dangers.htm). Both are found in the last 2 paragraphs on this page (Nov 8, 2004). Even the CACM had an entire issue devoted to end-user computing a couple months ago.
- As to your other point.... If I had an account, then I would not be the "phantom avenger for software engineering", I would have to be the "phantom avenger for software engineering with an account on Wikipedia", but that that would be someone else entirely different.
Scripting is big business
The most widely used programming language in the world is Visual Basic. It is used for all sorts of scripting work and custom programming. Spreadsheet templates and scripts encompass as much development every year as all other applications. That is reality.
User programming is a huge third branch of software.
This matters, because computer software is used on software engineering as the main definition of software. (Kenny did that). Anyways, applications tend to be written by SE people while users software is written by everybody. This distinction must be made clearly to define SE properly. If this page does not want to define all software, then the SE page must use some other page as the main definition.
- Regardless, I've never seen anyone separate application and user software into two different categories. I did not remove the distinction between them, I merely noted the subdivision. If you disagree, please provide references, and remember that this is not a place for "original research". --Shallot 15:08, 5 Aug 2004 (UTC)
- There is a lot of data about this on the web. There are 630,000 full time software engineers in the U.S. who spend most of their time doing application and system software. There are closer to 2,500,000 programmers in the U.S. who do a lot of programming. The number of compilers sold is close to the number of programmers. There are sites like data (http://www.sbtdc.org/pdf/software.pdf) which break down the application programming. Sales data on compilers gives a better count of actual programmers.
- Probably the most common mistake in understanding the software, is to mix it up with the software industry, which gets paid to make software. Most software is written by scientists, businessmen, hobbyists, and others who do not get paid for it directly. But, ignorance has never been accepted as justification for poor scholarship.
- That document from sbtdc.org has three categories, but they are "systems software", "application tools" and "applications software". The first is fairly obvious, the third is described as "includes products such as word-processing and spreadsheet programs", while the second is composed of "data access and retrieval, data management, data manipulation, and program design and development software". I don't see from this description that they mean templates, macros, simulations, scripts and filters; rather, it sounds to me like they delineate "mainstream" application software from more specialized applications. No? --Shallot 20:22, 5 Aug 2004 (UTC)
- You are following the usual for-pay industry = software metaphor. That document describes the for-pay industry well. To see the rest of the programmers, you only need to look around at everyone else who programs. That means counting compilers shipped and so on.
- The U.S. Bureau of labor statistics recognizes three different categories of software developmers as described in the following table. See BLS home page (http://www.bls.gov/oes/2002/oes_nat.htm) for more details. Note that these are the "full time" software developers. There are even more who have other job titles, but do software development anyway. This data is duplicated in Software engineering demographics.
CODE TITLE Number in 2002 15-1021 Computer Programmers 457,320 15-1031 Computer Software Engineers, Applications 356,760 15-1032 Computer Software Engineers, Systems Software 255,040
- According to the the U.S. government, the largest group of programmers does not work on either applications or systems software. They work on a wide variety of tasks, including scripting, customization, scientific simulations, etc.
- Hmm. But couldn't this statistic above simply mean that the largest group of people simply have broad, generic job descriptions? In other words, perhaps they don't actually spend all of their time doing applications or system software, but still do various things which also include both applications and/or system components?
- And even so, I'm still wary of separating a subcategory of application software and/or of data into a parallel category without an exact reference. Dunno. --Shallot 23:36, 7 Aug 2004 (UTC)
- To some extent you are correct. However, those programmers claim not to be doing applications or systems. And there are millions of spreadsheet writers. The number of users who customize their own software is huge. Visual basic has more lines of source code than any other language in history.
- To my mind, the real question is what kind of proof do you want. Here is a link to an editorial in IEEE Software that worries about end-user programmers link (http://www.computer.org/software/homepage/2004/July-Aug/dangers.htm). There are lots of articles out there like this. There is plenty of people who count end-user programmers (especially Microsoft). What do you want, a text-book reference?
- I just googled "spreadsheet programmers" and one of the top links was link (http://www.stsc.hill.af.mil/crosstalk/2004/06/0406Cook.html). The first paragraph says "It is predicted that by 2005 in the United States alone there will be 55 million enduser programmers compared to 2.75 million professional programmers".