How to Select Software Developer?

First, it requires as much work on your part than on theirs. You must define on paper what you need and when. What are the limitations and the performances expected. Break the project into many smaller objectives and make sure that you "own" all of the results at the end of every objective. This way, if you are not happy with the results, you can terminate the contract and continue with somebody else. The documentation is part of the performance. It is as important as the code itself. If the programmer tells you that he adds the comments later and write all the user doc when the program is done, don't use him. You will be forever his slave or will never get proper documentation. Ask for references and examples of work previously done. You should ask an independent programmer to help you evaluate the candidates. There are more incompetent programmers than competent ones out there. Often the worst ones are very good at using buzz words that make you believe that they know what they are talking about.

Finally, the programmer should understand the product you are working on and have done similar things before with the language that you intent to use. Somebody has to pay for the training of a "first time" even when executed by an experienced programmer.

First advice: start from standard software that is as close as possible to your requirements. Talk to existing users to see whether they have had issues with the sorts of extensions that you want (and who did the work for them if they did not do it themselves). Then consult the vendor as to their view of customisation suppliers. It's usually worth trying the unmodified software first to get an idea of any "hidden parameters" you will need to consider.

Second advice: the details of the specification are both as critical and at least as difficult as the software development itself; if you haven't done this before, you are likely to make assumptions that don't quite work out in practice. This is one of the reasons why software always costs more than expected.

Regarding additional selection criteria: integrity first - always. After that, communications skills. If you and the prospective developer have those, developer will not offer anything he cannot achieve (but he will almost certainly over-run, so ask for his track record on this and talk to previous clients).

I don't think you have given enough additional detail to say more.

TJSHAH:

Use the advise of posts #1 & #2 and put into www.rentacoder.com/RentACoder/default.asp Peruse other bid requests then take the plunge.

Well, that is really hard question! In principle good programer could program anything, provided it is adequately specified beforehand by YOU. Since on every good programer there is 99 who dont know what they are doing or have made one program and then are adapting it and making other programas from it, and everyone can brag as much they want while You are not able to check this, best way is to try to give small program that has all elements Your application would have (GUI, Files/Database writting and drawing on screen, printing) to prospective developer or developers as the assignement for the test. If You are satisfied that this first assignement work, that hire the programmer. For initiall checkup, You could ask for proofs that candidate was working where this person claim to have worked so far. I am 34 years in programming and have learned over 20 different programing languages, some of them in severall generations and from severall vendors on at least two platforms, and I still keep my employment notices so I can prove that I am telling the truth. So, if You could define the skeleton of Your application, the rest is easy: just fill in formulas. But it is You who should define data structures for database/files, and it is You who should know what is expected as result of program IN DETAILS. I could offer myself to work for You, but I have definite plans to make program that would be usable for design of any kind of devices on computer and simulating their work just like prototype would work in real life application, test them under every imaginable (but possible) condition and superoptimize them to be most efective and as cheap to produce as possible........ Anyhow, I also have definite opinion on some of >>fashionable<< programing languages, so maybe we would not find it possible to agree on this. Solution would be that I see what You would like to do, and if I think I could do it, I would acceppt assignement for fixed fee and reasonable time scale for finishing it, if You let me use my prefered programming language, of course. Othervise, if I see it is close to my own idea, we could make coproduction and I would program it while You would finance it, and where we would try to use best ideas out of all we both have.........

If You are interested, we could continue this conversation in private.