Re: Changes to Euphoria
ken mortenson wrote:
>
> Thoughts inspired by Jason and Jeremy's discussion...
>
> Things in life are often a trade off...
>
> Programmers develop habits as well as use idioms of a particular language.
> Working with a small team of programmers on a huge project I can often tell
> which member of my team wrote a particular section of code by how it's
> written. If I got my way, my ability to do this would lessen which is a loss.
>
> So what am I looking to get to compensate for this loss?
>
> I'm looking for a cleaner code that is easier to read with understanding.
> The quicker you get comprehension the less you are stressed. Making the
> programmers job easier. Idioms shared make comprehension easier too.
>
> You can achieve much of this by the minimalist approach.
>
> Take GOTO's verses structured code as a first example. GOTO allows a
> programmer to express themselves very individually (making it easier for
> me to tell who wrote the code) while structured code isn't so individual.
> The structure tends to enforce a likely similar looking code. And it's
> generally quicker to comprehend (especially in a long block of code.)
> Structures have a beginning so you know you're in a structure. Even if the
> code scrolls of the end of the page, you already know the eventuality.
>
> With GOTO all you see at the beginning of a loop is a label. That label
> doesn't tell you, you are about to enter a loop. Your comprehension will
> require you to examine more code.
>
> You could also minimize by taken out the structure, just use GOTO's
> everywhere. It's largely a matter of taste, but I prefer using structured
> code. YMMV.
>
> I'd like to talk about loops. You may think I'm just in love with my idea
> (which I admit is partly true) but there are good reasons for my proposal.
> I don't believe any on this forum will have any trouble judging the merits
> or speaking their mind about it. Some may even oppose an idea not for the
> idea itself but because of who proposes it (which I think is a rather sad
> thing.)
>
> It's been noted that FOR loops are an addition to Euphoria which originally
> had just while loops. I've used FOR loops so much that I don't give it a
> thought usually about whether they should be included or not, but thinking
> minimally (don't include what you don't need) I realize there's a good
> reason for not having them. Let me start with the while loop and get back
> to the for.
>
> while (1)
> ...
> end while
>
> I would claim that quicker comprehension occurs with this construct...
>
> Loop
> ...
> End Loop
>
> Both are infinite loops. Yes, it's not much quicker to comprehend, we're
> not dolts after all, but after examining thousands of lines of code, by
> 3am I betcha that it becomes a bit more significant.
>
> I would also claim (without support, just my INTJ assertiveness again) that
> comprehension is better when Loop is used to describe Loop, rather than any
> other _____ is a loop. Call me crazy.
>
> Now, how about getting out of the loop? Some language provide a plethora
> of ways and are thinking of new ways all the time no doubt (programmers are
> such a creative bunch after all.) The fact that sometimes you need to get
> out of loops and the beginning, the end, the middle has led to repeat, until,
> etc. But for a minimalist the answer is right there, IF. You need something
> like EXIT to be used with the IF, but IF eliminates the need for any
> additional unneeded keywords because you can put the IF test anywhere in the
> loop; beginning, middle and end. Exit tells you immediately that you are
> leaving the loop you are currently in. Exit 12, on the other hand, is an
> offramp in New Jersey I think!
>
> What about FOR. What does FOR add to the game that LOOP doesn't give you
> already? FOR gives you an INDEX. What a great idea. But perhaps
> if a minimalist were trying to add an implicit INDEX he or she might have
> realized you don't need to add FOR at all!
>
> Loop
> ...
> End Loop
>
> There it is, did you see it? Looks like anticomprehesion time, doesn't it?
> Never fear, I'll get back to that in a moment.
>
> FOR gives you more than just INDEX. You can tell where to start, where to
> end and how much to increment by.
>
> For LOOP to have a start you'd have to do something like this...
>
> Loop from 5
>
> End Loop
>
> Well why not? This also gives us something else...
>
> Loop
> ...
> End Loop
>
> ...under the covers can become...
>
> :Loop
> ...
> GOTO Loop
>
> And now quicker comprehension because the lack of from indicates no INDEX
> at all. It does however require the idiom Loop from 1, but I don't think
> that's such a big deal. Idioms foster comprehension too.
>
> Other languages use STEP where Euphoria uses BY... I really like BY. I
> don't know why, I just do. So we keep it...
>
> Loop from 10 by 5
>
> As long a we understand the INDEX is implicit (with only Loop for loops that
> comprehension will come rathe quickly. We don't have any other type of loop
> to cause confusion.)
>
> But how about the end of the loop? With FOR I specify the end in the
> declaration. We minimalist don't need it. We already have IF (cond) Exit.
>
> Again I would restate that use of a single loop construct would make
> comprehension quicker and would be reinforced with each subsequent use.
>
> I stated my case for the two issues I've brought up before (somebody got
> a fire extinguisher handy???)
>
> Am I in love with my ideas. Sure, why not? Can I live with rejection.
> Yeah, I suppose. I'll just go into a corner and sulk for awhile.... 
>
> The biggest problem is that taken up my proposal requires change. Change
> is not easy. No, better to say, change is hard. One of the really annoying
> things about me is that time often proves me right; by then the sun has
> cooled, the earth has left it's orbit and the guys I was fighting with (in
> the friendly fashion that programmers do) has left the company.
>
> But it's not about me. So please consider the merits regardless of the
> source.
>
> I would also argue that a minimalist approach makes for a better compiler
> or interpreter but we'd all agree on that, so where's the fun?
Frankly, your approach reminds me of Fortran (any flavour).
Its superiority in some specific areas of programming has survived programming
concepts and hardware changes. Fortran compilers have been the first to introduce
some advanced features like statistical reshuffling of code to optimise its
speed, etc.
Great and granted.
So you'd get the do i=1,n,step ... end loop, the arithmetic 3 way if and a few
other minimalist thingies.
But if I found Fortran comfortable to program with, I'd use Fortran, nand I
don't. I even downloaded OpenWatcom F77 compiler, just to try my hand. Good for
stuff where speed is more importan than maintainability. But I'd say hardly
better.
CChris
|
Not Categorized, Please Help
|
|
