Discussion Locked

The right long term devleopment language

0 helpful
14 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

My question to you is just that. Which development language and platform should corporates be reviewing. With the availablity of Developers at a premium the market availability of VB programers over say Delphi, C an Java its hard to offer any alternative for a _solid_ long term strategy within any organisation faced with the option of system renewal and internal implementation.
In broad terms Java is slow(comparativley) and still young in the market place, although I accept there is a very wide tool set.
Delphi does not have the market penetration and the big boys behind it to realy present a serious option.
C(and its variations like C++ and C#) presents a good alround option but, IMHO, the basis of the technology is seriously outdate - OK I havent used C# and I would hope its rocks but I would put my hopes on it plus its way to early to say.

So whats your view on the matter.

Iain

A mind stretched to accomodate a new idea never shrinks back it its original state.

14 replies

by tim leedy

Unfortunately there is no easy answer for your question. Your question raises other questions:
1. How large are the applications that the corporation will be developing? Are they strictly internal applications or will they be used by outside clients?
2. Does the language have lasting power in the market place or will it be gone in five years?
3. Does the language provide the functionality that is needed for the projects being developed today versus the projects being developed in five years?

These questions should be answered before you can make the correct decision. For example, I worked at a company that invested in Gupta SQL Windows as a development language years ago. At that time, SQL Windows had a large share of the marketplace. But, over the years SQL Windows use in the marketplace dropped and the language was not keeping up with new advances in technology. So the company was forced to re-write the applications in another language.

You might also want to think of a multiple language solution. That way, depending on the project, your company can choose the best language to complete the task.

by wayne m.

Pick any proven language that fits your needs. None of them are going away.

A short list of possible languages would include C, C++, COBOL, VB, and Java. All of these have a sufficiently large program base to ensure they will continue to exist.

It is hard to predict the staying power of new languages such as C#. They may catch on or they may be relegated to obscurity. Who can tell?

Pick long term languages like you pick long term stocks. Blue chips are always safer then IPOs/new languages.

by hengis

I am a techie at heart and to think that COBOL is still a valid corporate choice for _new_ systems would concern me. Fair enough keep the language up to date for legacy integration but I would have thought you would want to go for a more 'modern' approach

The use of the word modern might be a bit unfair on COBOL but my understanding was it is purely legacy, am I right?

While I am on, would you consider Pascal/Delphi as a valid platform, it has been arround for a while and it does have good support tool wise.
Full disclosure: I have used delphi in teh past so I an biased towards it

by wayne m.

If you are working in an environment dominated by COBOL programs, then you should develop your new applications in COBOL. Why would you want to support 99 COBOL apps and 1 Java app?

COBOL has continued to evolve. It how provides support for HTML and web page generation.

I am not a COBOL developer, but I work with enough of them to know that new COBOL applications continue to be developed, primarily because of legacy applications also written in COBOL.

This leads back to my originalpoint. Any language with a large amount of legacy code will continue to live on. The existing code will need to be maintained. New applications related to the existing code will need to be developed. Finally, the legacy code base will exert pressure on the language itself to evelove and adapt to new technologies.

by al macintyre

EVERY LANGUAGE IS LEGACY in the eyes of the person learning the next language that comes along to replace it.

EVERY APPLICATION IS LEGACY in the eyes of the people learning the next thing that comes along.

I happen to program in an old versionof RPG, but the issue for a company should be not what is the latest & greatest stuff, but the productivity of the people using the tools & the cost to the company to get the job done.

How many different languages can one person be expected to beproficient in, and does it cost a company any more to be functioning with a few languages as opposed to many?

So long as the language & its tools are supported by vendors, and there are plenty of classes available in them, and books on how to usethem, and the language is capable of doing just about anything that any company would want to do with any language, which is true for both RPG & COBOL then I think that is a key question.

Now if the only vendors that support some language appear to be at risk of going out of business, or education is not readily available, or if the language is not capable of doing the latest web functions or data base manipulation, then that might be an argument in favor of moving on to a language which does not have those failings.

by debuggist

It depends on your environment. If you have legacy code, it is hard to turn your back on that investment. On the other hand, you have to use what developers are willing to use.

Before you knock Java for being slow, you need to do your research. Sun (and others) have made and continue to make significant performance improvements to server and client-side virtual machines.

The best suggestion I can give is choose something that has a syntax similar to C and, if possible, a language that supports object-oriented programming to some degree. This will insure that you will always be able to find developers that can quickly understand your code.

It would be interesting to see the results of your search. Could you post them in this discussion or start another discussion if its months later.

Good luck on your search.

-Doug

by davidandgoliath

...that there may be no right single answer. As standards develop further and languages are able to talk to each other through interfaces such as COM, sometimes companies are forced to use what resources they have. They may develop a system with parts in VB and C++ for example. While this can be a maintenance headache sometimes it is forced by the realities of business needs.

I think the speed issue is overrated in many cases. So what if C++ runs 20% faster than Java (I'm just using these number as an example), in most applications database access is orders of magnitude slower that the speed at which today's processor execute instructions in memory. Sure there are critical parts of applications that need special consideration but don't get hung up on speed and develop an overly complicated system.

by samuel c.

C++ is better than C and less dangerous.
There is no alternative to COBOL for Business
Applications. Also Ask What Database you are
going to use? VB is good for light things
in Windows Environment. Java for Internet
Environment. The Fight With Microsoft has not done much good to the Industry, and still
leaves to many Open Questions.

by al macintyre

If you have seen the latest bench marks that IBM ads are touting for their model 840, you know that any language will get faster in the future.

I am simply amazed that IBM Universal Data Base was certified for Windows 2000 before Microsoft's own SQL server that does the same work. Of course who is certified first is not the main issue when comparing Oracle & other vendor tools, The issue is which do the work productively & hand shake well with other vendor products, so that we can write code that works across many applications using different kinds of protocols with minimal hassles.

by dr dij

C AND C++ seem to me to be low level languages to some degree. Case in point: I had an app compiler that took my 2 paragraph Basic program to read a database file, input and manipulate data, write back records. This cross compiled to 2 pages of 'C' calls to update the Informix database.
Another case in point: The developers of MAS-90 spent lots of money developing a version of their accounting package in C or C++. It never made it to market because they could never get it to work. They reverted back to the Providex basic.

Sure it's great for motor controlling or low level stuff but I don't think you'll ever find a major ERP system written in it. I could be wrong- are there any out there? but each statement doesn't do as much asa high level language so you're less productive unless you do mostly calls to prepackaged routines.

by albion1

COBOL development in the mainframe environment has not slowed down at all. Despite all the hype, for large scale operations a mainframe is still the best option -- client/server technology just isn't there yet. On the mainframe, in business, COBOL rules.

It is also a very good programming language: It is not difficult to learn technically. Even a non-programmer could read through a pgm and figure out what is going on. Unlike C/C++, there is very little impetus for propeller-head, techno-macho pgm'ing. If you are turned on by bit-twiddling, you have to go out of your way to do it.

Just about everything in COBOL is backward compatible to the year zot. This means that the pgm written in 1968, unless it contains any of the relatively few obsolete lang. elements, will compile in the newest compilers. Try that with Access 2 or the original VB.

It has evolved dramatically, especially over the last few years -- there is a new Standard out there; does anybody know that! And you can develop COBOL applications to run in UNIX and client/server environments. If you are really sado-masochistic you can even write Object Oriented COBOL...

There are also a lot of experienced COBOL programmers and they don't (unfortunately) cost as much as programmers in C/C++, Delphi, VB, Java, etc.

by bresnik

Having developed products in C/C++ [some legacy non-OOP and new C++] and in Delphi, my choice is Delphi for speed of development, speed of compilation, and even comaprable execution speed. However, until Kylix is released, I'm also learning Java [using JBuilder providing the best IDE to date].
Long term... it's easy enough to switch from any decent object-oriented language to another. As long as our clients feel Delphi can do the job, we'll stay with it!

by samuel c.

If Delphi was a standard language then OK.
It is (if I am not mistaken) one of many
Rapid Application Developers.
The Point of this discussion is that after
Half a Century of Programming there is not
an ideal Programming Language.
And let me add there is not an Ideal Operating System too...

This discussion is closed to new replies.