quote:Original post by HenryApe
Actually, a number of studies has shown bslayerw to be correct about Java development being quicker and less buggy than C++ development

I suspect that you could find a number of studies that show the opposite if you so wished. The point being that studies tend to prove whatever they are weighted to prove, and I've certainly seen several bogus studies being taken very seriously. People believe what they want to believe, and then gather "evidence" to support that belief.

quote:
but I guess that is also the reason why you decided to answer him with a silly Dilbert flame rather than try to prove him wrong.

You have an odd idea of what constitutes a flame. It was actually a joke, with some measure of truth behind it. A flame tends to contain some form of personal attack. The "PHB radar" is highly sensitive to brand marketing. Java's rise in popularity has been accompanied by plenty of that. There are very few technologies which succeed on the basis of technical merit alone, hence my point that the PHB radar is likely to be a bigger influence on a technology's success. The PHB radar acts orthogonally to actual technical merit.

[edited by - SabreMan on August 30, 2002 5:09:17 AM]

quote:I suspect that you could find a number of studies that show the opposite if you so wished.

Suspicions are always useful when you want to discredit someone but can not actually prove them wrong. I am pretty sure that you can not find a number of studies from reputable institutions that show C++ development to be quicker and less buggy than Java. However, even Microsoft refers to a productivity study from IDC that compares Java to C++ and concludes that Java development is much quicker than C++ development (I could only find that page in Google''s cache as Microsoft seems to have reorganized their site).

quote:People believe what they want to believe, and then gather "evidence" to support that belief.

Yeah, and it is interesting to see how you are trying to dismiss the software metrics field as subjective because it does not agree with your beliefs.

quote:If you were the decision-maker for a well established company developing a graphics-intensive 3D game where framerate will be an important consideration and there may be many entities running/flying around, which language would you wise to choose (not talking personal favorites)?

Well, if the company is established enough to do console development, then C/C++ is currently their only choice. Hopefully this will change in the future, but right now that is a showstopper for other languages.

quote:Original post by HenryApe
Suspicions are always useful when you want to discredit someone but can not actually prove them wrong.

My comments haven''t been about proving anyone wrong - they are not mutually exclusive with what else has been said. Don''t forget that I said the reasons for adopting a technology are often orthogonal to the merits of that technology. IOW, what I''m saying is not that Java fails to impart the benefits you''ve described, but that I''m not convinced those benefits are the main influencing factor in adopting Java, or necessarily that they should be. There''s no "attempt to discredit" there.

quote:
I am pretty sure that you can not find a number of studies from reputable institutions that show C++ development to be quicker and less buggy than Java.

Maybe. Maybe not. I''m not that interested. FWIW, my experience leads me to believe that Java does impart a slight advantage in speed of development, but not "bugginess". Greater factors in those areas include the abilities of the team members, the organisational capability and the procedures surrounding the project. Perhaps the biggest issue for C++ shops is that it takes more intelligent developers to gain similar results to Java, but conversely that it allows greater freedom of expression so that such developers may achieve "better" results. Part of the *real* problem is simply that there are so few good developers around.

Another point worth making is that there are other technologies around that are better at what you are claiming Java to be good at: Python and Ruby spring to mind. Hell, if we did go with the *best* tools for the job, C++ would never have gained such a strong position, and Java sure as hell wouldn''t get a look-in.

quote:
Yeah, and it is interesting to see how you are trying to dismiss the software metrics field as subjective because it does not agree with your beliefs.

It''s even more interesting to see how you project various beliefs onto me so that you can disagree with them. I don''t hold the beliefs you seem to think I do. FWIW, I am consistent in my skepticism of software metrics. It''s not that I don''t think there are worthwhile analyses out there, but its always difficult to assess the credibility of the source. In-house studies are often better, as they are able to take into account local factors. Generalised studies often have an "everything else being equal" subtext, and often everything else is not equal.

And while I''m on it, I might aswell point out that the most recent such study I''ve conducted on behalf of my client has concluded that they would be better off migrating to Java. Make of that what you will (although you didn''t seem to require an invitation to do just that with my earlier posts).