are no more accomplished (and maybe less) than American/European C programmers. So, IMO, insistence on Ada, and the time spent to absorb it, would not be a major added expense, and the improvement in code quality should more than pay for it. Use of Ada would be an advantage, because it would avoid the prevalent mis-application of C. I would recommend Ada rather than Modula because it is almost as universally available as is C, due to inclusion in gcc.
--
[mail]: Chuck F (cbfalconer at maineline dot net)
[page]:
Try the download section.
** Posted from http://www.teranews.com **
The problem with the very large number of programmers "proficient in C" in India, is that for a large percentage, the proficiency is not much more than writing "Hello World" programs.
Such proficiency level can be obtained with short crash courses, but in order to use such programmers in any real project, a competent programming manager is required for each 2-4 such inexperienced programmers. In the worst case, the experienced programmers spends all his/her time managing the inexperienced programmers, while he/she would have produced the same code alone in the same or less time.
A country ends up into this kind of situation, if it tries to bootstrap the IT industry (or in fact any other advanced industry). A much better approach is to include only one inexperienced programmer into a team of programmers with at least moderately competence.
On a country level, this means it takes a lot longer to bootstrap the industry, but the results are better in the end.
It seems that some Indian companies have realized this, since they are buying IT companies in Europe. When asked why they are doing this, while they have a huge number of programmers in India, the Indian owner of one such company admitted in a TV interview that in India, they have the required quantity but not the sufficient quality of programmers and designers.
Paul
The same can be said for just about any other programming language, with the exception of programming languages that are mainly used for safety critical applications.
No disagreement there. Another concern is mentality. Too many programmers are satisfied when their code appears to work, even when they do not fully understand why it works. I have encountered too many cases where the programmer didn't fully understand the semantics the constructs of the programming language he was using. Some won't use a language construct before the fully understand its implications, others (in my experience too many) just muddle on until the programs appears to work (edit&continue debugging must have been invented for those people). IMHO with the latter attitude you cannot write safety critical programs, no matter what programming language you use. One can write crap in any language, only some languages makes it particularly easy.
I think too much importance is given to the used language. Good programmers have been through a number of languages and if needed could quickly adjust to any other one; those really good would have settled to a certain language for a reason which might prove worth enough the respect to go lax on the language requirement (given that the available tools allow that, that is).
My observations are the same. I guess things will not improve any time soon, most people are not up to the power and complexity of the technology of the day - but have wide access to most of them (not that I am preaching to limit this access, this would be worse). Which explains to a large extent why the current software picture is as messy as it is - with some rare exceptions.
Dimiter
------------------------------------------------------ Dimiter Popoff Transgalactic Instruments
------------------------------------------------------
Original message:
... snip ...
That depends on the programmer. Good programmers can cope, or alternatively are good programmers because they have had to cope in the past.
Today we see far too many who can use nothing but C (sometimes C++, which makes it worse), and that poorly. They have no idea of the fundamentals that drive the language, in fact assembly is another world for them. Suggest Modula, Lisp, Ada, Pascal, Cobol, Fortran, Whatsis, and they recall in horror from the unknown.
--
[mail]: Chuck F (cbfalconer at maineline dot net)
[page]:
Try the download section.
** Posted from http://www.teranews.com **
Much better--we'll make a manager of you yet! There were some incentives and other programs initiated. I'm not sure how well they worked. One of the programs was to send them to the US for a year--in exchange for which they were had to commit to staying with the company for some period of time. As a US manager said: "I thought the British tried that--didn't work for them."
~Dave T~
Well, this was a decade or so ago that this started. And would have meant training the non-Indian programmers (at the automotive supplier and at the very large auto company) in Ada at a time when the auto company's worldwide purchasing was insisting on "commonality" in order to reduce costs. They believed (still do AFAIK) that since C was so common, specifying its' use would allow them to purchase from anyone, increasing competition and reducing cost.
~Dave T~
Here's an interesting article in this month's Crosstalk on "Lean Software Development". There are some tables showing the relative effects of various software approaches on productivity, quality etc. Outsourcing is the only one show with net reduction in the "ilities" compared to the structured methods of the 70's. I can easily believe it, based on some of the anecdotes I've heard from colleagues who've managed such efforts.
=A0
=A0
of =A0
Many thanks for that fascinating reference. I wonder if the outsourcing crowd will reconsider when they see that 20% quality impact figure (that's down from the 100% baseline!) for geographic outsourcing. Yowsa!
BTW, every in-country programmer in the western world ought to send this information to their higher-ups.
Jim Sutton's book, where he goes into the details of the approach, is really good, if you're interested in these ideas.
Yes, thanks, I'll definitely be getting it.
An interesting related point is that of the implementation of type extension and dispatching in Ada95. The designers of Ada95 didn't go to the extent of implementing the object.method() syntax that now exists in Ada05. Rather, subprograms used the concept of a "controlling parameter" in their signature which controlled the version of that subprogram that was used for the particular object. When I first came across that mechanism it confused me, but once I understood it, it taught me massive amounts about how all this stuff fitted together. In C++/Java you (well, I) tend to take that for granted.
Yes, I know. There may be interesting times ahead ;-)
Did you mean "Lean Software Strategies" ?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.