Re: v2.5 Opens exw files wayyyy too slow
- Posted by Georg Wrede <georg at iki.fi> Dec 24, 2004
- 568 views
Various authors: > To see how professionals like Georg do it, see: > http://databrook.com/users/irvm/photo.jpg Damn, I've been trying to download the picture, but it's been continuously donw for the last 6 hours. :( > We could all sleep in caves still too :P > You can't live in the past and use modern computers. > The old paper ways Sure don't beat Edit -> Run -> Edit -> Run > With all due respect Mr. Wrede, sir: you’re so full of s... > you’re unsanitary. Thank you all for your attention. From your responses it seems you belong to the lucky ones who are not terminally spoiled by edit-run-edit-run style. I see every week programmers (both young and old) who do it that way, and some of them really do not get it when told that programming is not the same as typing code. Ah, well, maybe they'll learn later. ----- Chris Bensler wrote: > If Euphoria would return all parser errors at once, what you > are saying might be more acceptible. I wouldn't be surprised if that feature appeared in Euphoria. But at the moment there may be more pressing things to do. > However, interpeted programming is not quite the same as compiled > programming. With compilers, you have extensive error reporting, > and a heavy duty debugger, and a degree in computer science to be > able to use them effectively. Hmm. Since Euphoria is an interpreted language... That you can compile Euphoria code doesn't change the fact that it is primarily an interpreted language. So you are exactly right. > You are right in that people should be planning before they code, > but I don't know anybody that plans mistakes. Yes, they _should_ plan before they code. Seen anybody do it? Given an assignment everybody runs to the keyboard. Once I visited a programming firm with a good track record. They told me their "secret": since no programmer wants to do planning, they decided to have them make prototypes. But to avoid prototype code ending in the final product, they decided to have them do the prototypes in another language. The programmers could choose their language freely, as long as it was not the Production language. The improvement in code quality was more than they had hoped for. Also, the total time for finalized projects was actually about the same as before. (The most surprised were the programmers.) > I should add: > Why is it do you think that almost every compiler (actually every > one I've ever seen) returns as many errors as it can, even if > those errors are cascaded? I agree with you. Who could do a major project in C++ with a compiler that pukes at the first error? > Also, why is that almost every compiler has a highly advanced > debuggger? Money. If Rob had money to hire a big staff, I'd bet we had an industrial strength debugger, a compiler that not only finds dozens of errors, but also would give state-of-the-art help and advice on them. Maybe even correct some of the most obvious mistakes? > The answer is: so that you don't have to spend so much time in > compilation. Granted nowadays with faster machines, compilation > time is not as big of a concern anymore. Some alternative ways to cut the time spent compiling in half: 1. Buy a machine twice as fast. 2. Have Rob jump somersaults and go through the needle, to really make the compiler scream. 3. Learn to touch type, so you can keep your eyes on the screen, so you catch your typos immediately. 4. Learn the (very small) grammar, and the most used idioms! There exist few languages with less to learn than Euphoria. 5. Draw a picture of your own data structures! And hang it above the monitor. Oh, and do all five, and you spend 1/(2**5) time compiling. Then you can stick a label on your cubicle: "I spend 96.9% less time compiling." -- Another Euphoric, since Nov. 18, 2004 --