Re: Standards/reasonable programming practices (&GPP)
In article <5np6t6$jds@cssun.mathcs.emory.edu>, "Igor I. Lisyansky" <fox@acdlab.garnet.ru> writes >Hello All! > >It's one of the great achievements of human thought encountered in this thread: > >> Yep! Bunch of Clueless )*(&^*^ers. My latest theory is that the >> undisplined youths you heard about on TV - their generation >> are now coming out of university from CS courses and creaing crap code. >> >> I'll have a job for live fixing their code and my ability will show >> as more of them appear an the overall quality of coders goes down. >> >> Welcome cool windoze coders with their fab graphics/icons, you'll >> keep me employed for life on a good salary as a top coder.... > >IMHO, it's not so plain. Maybe, You think that command-line tools make a person >a good coder >and the modern IDEs do the reverse job? Sorry, the basic principles of >programming (like >the famous Dijkstra's theorem) doesn't depend on the IDE. It's quite possible to >write a >good code in Delphi and abominable spam in GNU C. >Also, I've never heard of colleges, universities and even CS courses where the >folks are >taught to write code bad code or use the Clipboard instead of procedures and >functions. >However, when the developer comes to the company, the managers usually require >to create >something ASAP, with no attention to such insignificant things like good >programming >practice, robustness, GUI logic etc., but with great attention to their own >prejudices in the >sphere of GUI. >The managers need manageable people, who can be easily manipulated and replaced >at any time. IMHO, good developers aren't usually folks of the kind. Very true, in my previous place I stayed 2 years after the management wanted to fire me...and even after telling the guy one below the MD and another director to F**K OFF! Why, because I'm good and they couldn't replace me. They replaced all my friends and about 40% of the company but not me... >You probably could see the announcements of the RAD tools allowing 'the >beginners to create >full- featured applications' (especially, the notorious SQL Windows). Maybe, the >developers could Exactly, users and managers think it is easy and a 5 minute job... >say this? In any environment, in any OS, the skills, knowledge, experience and >culture are >required to write something other than 'Hello, World' with multiple bugs. when it is not. Just because a user and run a few queries and simple reports in a GUI based report writers does not mean they can replace proper programmers as they are beginning to find out. Yes that report has a processing loop in it which annot be repplciated by one SQL statement... >It's common practice to criticize Microsoft and its products. However, who >prohibits to buy and use >X Window based systems? Nobody, it's also a managerial issue, known as seeking >for cheese for free. Or seeking a complete system coded in a day.. >To some points, the vendors of UNIX and UNIX systems aren't also innocent. When >I buy a copy >of Borland C++ or Delphi, I get the fully functional development tool with all >the functionality of >the operating system available. However, I have some experience with >SPARCompiler C++. >We paid $2000 for it, and what we got? The plain command line compiler, without >even the manuals >on OpenLook objects, without programmatic interfaces to XIL and XGL libraries, >with MakeTool >dumping core instead of running. Maybe, this practice could increase the amount >of applications >written for Solaris/OpenWindows and thus the interest of users to this system? >Buttons and icons allow a serious programmer to concentrate on the program and >its objects rather >than on the syntax of makefiles and ways of finding the necessary line in the >text editor. Agreed. >To the bad programmer, they give a chance to eventually create something >formally resembling > an executable. But, IMHO, there's no reason to confuse such activities with the >Software Engineering. This is the real problem, c00l dudes how think they can program because they can put together a simple program which does half of what it needs to, hs no error checking and would run like a dog on a real sized data set (E.g. 200,000 rows). >In any case, I see many folks claiming Linux the best operating system. My own >experience >confirms it. Why? Linux was created by the community of developers rather than >by spam-sellers. Eactly.. >So, IMHO, Windows has done its very useful job. It made the customers to >purchase the systems >capable of running normal operating systems. DOS couldn't do it. Ha! Ha! :-> >So, in my opinion, the summary is the following. The developers would turn to >good programming >practice only when the users will definitely request for good software. Also, >the developers are at >least not only folks to be claimed guilty here. > Yep! I think the really good devlopers will have fun fixed all the badly written code which I think will appear real soon now... > > Igor I. >Lisyansky, Software Engineer > -- David Williams