> Boohoo. 10 seconds? When was the last time Word started up in under 10 > seconds on any machine with less then 2 Gb RAM? It's the operational > speed ONCE RUNNING that is important, not the startup time really. > It's a small price to pay for the privilege of not having to port > (recompile) the code to each platform every time you make a change.
Well, I don't remember the details exactly - it was several years ago (I'm not even sure how many years!). I just do remember such a delay for a very small console based app - that's why I brought that example. > It drove me mad too to begin with, and in fact you CAN avoid the > layout managers and position everything manually. But eventually > you'll realise that doing so is SO internationalisation-unfriendly, > since the programmer choosing sizes (unless he or she chooses way too > big just to make sure everything will fit, also making it look ugly) > won't be the right size for all languages. Once you realise the power > of layout managers, you won't look back; just the same as I didn't > regret learning C++ / MFC when I thought it was pants. It's part of > the learning curve of Java, just like learning to use std:: methods > will be part of the learning curve for me with C++. You can't expect > to get everything down pat the first week... that's arrogance. Hmmm, I admit I never put as much time into any language as I've put into C++ (including Java) so maybe my comments don't count. Not to say that Java is bad, but currently any time I put into learning it doesn't pay me back (again not because Java is bad, because my field is not what Java would fit into), so I have a learning limitation as well. > You'll also find the event model frustrating, not being able to force > a repaint of something effectively when you want; you simply have to > break the old habits and learn the ones provided, since they ARE > powerful when used correctly. For example, in Java when you run some > lengthy code from a button press, the button will appear to remain > stuck in until the routine returns on your first 100 attempts. > Eventually you will find out that by using a special technique > provided in SwingUtilities, this can be avoided; you can schedule that > lengthy code to run as soon as the UI has completed redrawing itself, > i.e., the event queue is empty. Well, IIRC the same is true with an MFC app. Of course you could use a multithreaded approach for such cases, but I don't remember if (or how good) does Java support multi-threading. > Of course it's slower - its interpreted - what did you expect? It's a > trade-off: see if you can get your C++ program working from a web page > under any respectable browser (IE, NS, Opera, Konqueror, etc.) without > writing multiple versions of the code to do so. And I'm not talking > just script functionality > here: I'm talking proper programs that run within themselves. Actually the speed is critical in my project. I'm getting involved in a team developing MySQL support for Persian. They are looking into an application which translated a whole MySQL database from Windows-1256 (and old and improper charset used for Persian) into UTF-8. So, there's a good amount of processing involved. Also, because there are already facilities available to do the conversion in C (like iconv), I don't wish to write them from scratch in any other language. And I can get a full blown app in C or C++ finished much faster than in Java. So I think it's not the right tool for my job. > First, almost everybody has the Sun JVM nowadays anyway. Can you > imagine someone having a PC that can't view Java applets? Well, back then I used VC++6, it required a JVM installation, since when I moved over from VC++6, I haven't installed JVM (other than temporarily to test some app, and then uninstalling it.) And the last time I visited a site which used Java applets... :-) They're very rare, or maybe that's just the sites that I frequent? > It's just not likely. And the JRE (runtimes only) is 9 Mb, a one-off > download; most people download more than that in their first week. > It's just excuses. Unfortunately, things are not this good here in Iran. The usual net connection speed is 33.6k here, and that's just the nominal speed. I have often seen people who will get excited if they can download more than a half of KB per seconds (yes, 0.5KB/s! not kidding.) And add it to the fact that most people haven't heard of download managers. So, if the app grows such big, effectively it won't be usable to many people. > Secondly, Sun's JVM is totally free. I don't know where you got the > idea from, perhaps some more MS propaganda, but go to java.sun.com and > follow the links - you don't have to pay a penny. When I said free, I meant free as in freedom, not as in "free beer". Sun's JVM license is not compatible with Free Software licenses, such as the GNU. I was not talking about the price (which is, of course, free.) > You know what? If people want to go on peddling the myths about Java, > trying to pretend it's no good and this, that and the other just to > avoid their comfortable little position where they think they have the > answers, go right > ahead: watch how little I care. I made a suggestion that I feel is > valid (despite being slightly outside your requested conditions), and > I stand by it. People can take it or leave it. I'm moving on... Well, I guess it's not the right tool for my job, at least. But, I do appreciate you suggesting it, and I find nothing wrong with anyone seeing Java as a good/bad development tool. At the moment, I just care about the job at hand! :-) Quite selfish, or what?! :-) ------------- Ehsan Akhgari Farda Technology (http://www.farda-tech.com/) List Owner: [EMAIL PROTECTED] [ Email: [EMAIL PROTECTED] ] [ WWW: http://www.beginthread.com/Ehsan ] Wishing that I could turn back time, so we could join our souls. _______________________________________________ msvc mailing list [EMAIL PROTECTED] See http://beginthread.com/mailman/listinfo/msvc_beginthread.com for subscription changes, and list archive.
