On Gotos and the long thread - some thoughts
C'est un message de format MIME en plusieurs parties.
------=_NextPart_000_0152_01C24D31.A9A38F40
charset="iso-8859-1"
Hi everyone !
I'm afraid some of us might miss the main issue while talking about =
the dreaded Goto.
(*) Since Euphoria's interpreter is not open source, the community =
does not have the final word on the features that will be added. RDS =
has.
So, IMO, the issue is not: "Should goto be added?". The community is =
divided on this, as the length of the thread shows. It is:"Will RDS (be =
persuaded to) add goto to Euphoria?".
The only thing the bickering suggests is a "Wait and see" attitude. =
The sad-looking statement on the future of "continue" shows it quite =
neatly, I think.
(*) is not completely true, since one can always buy the source and =
produce one's own interpreter. This is likely to require a high-volume =
investment in time for understanding the source, properly modifying it =
and insufficiently testing it (never forget tha Murphy was an optimist). =
And the issues of portability, compatibility, licensing etc etc start =
coming up.
My personal stance on *goto*: it isn't either good or bad. Coding is =
good or bad, the evauating criteria being readibility and =
maintainability of source. I coded quite a few things in Basic and never =
needed goto, so it wouldn't bother me.
On structured programming.
May I attempt to give a definition?
A project is a sequence of statements that are expected to perform =
several things (the functional specs). The more a coding scheme can be =
matched again these specs, the more structured it is. A language is said =
to be structured if it encourages and allows structured coding.
So, you can code in a structured way even in assembly language. But =
the "encourages" criterium is not met, so we generally consider it as an =
unstructured language.
Pascal, for instance, is so rigid that you are forced to code in a =
structured way. But sometimes the structure is not the one you need. C =
and Euphoria seem to stand somehow in an acceptable compromise.
Goto and structured coding.
True, Kat, any flow control instruction is a goto somehow, so all of =
them can be replaced by a goto on a lexical standpoint. But not on a =
semantical standpoint.
Consider the two following cases, lexically equivalent:
A: According to the result of some test, one of two or more program =
blocks are to be executed. They just reflect an alternative, and are =
equivalent in programmer's mind. The if/then/else.. construct (or =
select/case) tranlates quite correctly the flow picture the programmer =
has in mind while coding.
B: In a special case, identified by the result of some test, things get =
out of the general case frame, which the main program block is about. =
Then, some form of *goto* (goto,exit,resume,whatever...) reflects more =
closely the picture in programmer's mind, even though he may have to use =
the if/then/else construct for the interpreter/compiler to output an =
executable that performs as expected.
So, the absence or presence of gotos is not even a sign of =
structuredness. A good language would allow a programmer to speak his =
mind freely while being understandable by an interpreter/compiler in the =
appropriate way, and by other programmers who will follow on the =
project. This looks like an utopia; and one may go some steps forward or =
backward in this direction. Euphoria has a more intuitive treatment of =
sequence (forward), but a limited treatment of loops and branching =
(backward).
Have a nice day!
CChris (don't want to steal Bensler's signature - apologies)
------=_NextPart_000_0152_01C24D31.A9A38F40
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Diso-8859-1">
<META content=3D"MSHTML 5.50.4616.200" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3DArial size=3D2> Hi everyone =
!</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> I'm afraid some of =
us might miss=20
the main issue while talking about the dreaded Goto.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> (*) Since Euphoria's =
interpreter is not open source, the community does not have the =
final word=20
on the features that will be added. RDS has.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2> So, IMO, the =
issue is not:=20
"Should goto be added?". The community is divided on this, as the length =
of the=20
thread shows. It is:"Will RDS (be persuaded to) add goto to=20
Euphoria?".</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> The only thing the =
bickering=20
suggests is a "Wait and see" attitude. The sad-looking statement on the =
future=20
of "continue" shows it quite neatly, I think.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> (*) is not =
completely true,=20
since one can always buy the source and produce one's own interpreter. =
This is=20
likely to require a high-volume investment in time for understanding the =
source,=20
properly modifying it and insufficiently testing it (never forget tha =
Murphy was=20
an optimist). And the issues of portability, compatibility, licensing =
etc etc=20
start coming up.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> My personal stance =
on *goto*: it=20
isn't either good or bad. Coding is good or bad, the evauating criteria =
being=20
readibility and maintainability of source. I coded quite a few things in =
Basic=20
and never needed goto, so it wouldn't bother me.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> On structured=20
programming.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> May I attempt to =
give a=20
definition?</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> A project is a =
sequence of=20
statements that are expected to perform several things (the functional =
specs).=20
The more a coding scheme can be matched again these specs, the more =
structured=20
it is. A language is said to be structured if it encourages and allows=20
structured coding.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2> So, you can code in =
a structured=20
way even in assembly language. But the "encourages" criterium is not =
met, so we=20
generally consider it as an unstructured language.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2> Pascal, for =
instance, is so=20
rigid that you are forced to code in a structured way. But sometimes the =
structure is not the one you need. C and Euphoria seem to stand =
somehow in=20
an acceptable compromise.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> Goto and structured=20
coding.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> True, Kat, any flow =
control=20
instruction is a goto somehow, so all of them can be replaced by a goto =
on a=20
lexical standpoint. But not on a semantical standpoint.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2> Consider the two =
following=20
cases, lexically equivalent:</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2>A: According to the result of some =
test, one of two=20
or more program blocks are to be executed. They just reflect an =
alternative, and=20
are equivalent in programmer's mind. The if/then/else.. construct (or=20
select/case) tranlates quite correctly the flow picture the programmer =
has in=20
mind while coding.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2>B: In a special case, identified by the =
result of=20
some test, things get out of the general case frame, which the main =
program=20
block is about. Then, some form of *goto* (goto,exit,resume,whatever...) =
reflects more closely the picture in programmer's mind, even though he =
may have=20
to use the if/then/else construct for the interpreter/compiler to output =
an=20
executable that performs as expected.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> So, the absence or =
presence of=20
gotos is not even a sign of structuredness. A good language would allow =
a=20
programmer to speak his mind freely while being understandable by an=20
interpreter/compiler in the appropriate way, and by other programmers =
who will=20
follow on the project. This looks like an utopia; and one may go some =
steps=20
forward or backward in this direction. Euphoria has a more intuitive =
treatment=20
of sequence (forward), but a limited treatment of loops and branching=20
(backward).</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> Have a nice =
day!</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2> CChris (don't want =
to steal=20
Bensler's signature - apologies)</FONT></DIV>
------=_NextPart_000_0152_01C24D31.A9A38F40--
|
Not Categorized, Please Help
|
|
