OpenEuphoria

Re: A controversial topic

• Posted by petelomax Feb 23, 2023
• 870 views

I found this the other day on the Nim board, and it has been kinda nagging at me and ringing round my head:

Of course I get where (s)he's coming from, although we do want to hear what people want, but when someone gets annoyed that way,
I personally find it deeply disturbing, pretty much the sign of a sociopath, and something that digs up distressing memories of times
when I have done exactly the same myself, which is "blame everyone else" for their(/my) own inadequacies, inability, and inaction.

Now that really was't meant with any malice, though I'm quite sure it will be taken that way, and obviously I hesitated to say
anything at all, but it started gnawing at me in a way that keeping quiet about it was just not healthy for me anymore.

It was also stopping me from saying this: I just want it to go on record that I think multiple includes are utterly ridiculous.

include dog.e as bill 
include dog.e as ben 

(plus no doubt include statements within a loop and other madness[1]) is supposedly somehow much better than

include dog.e 
sequence my_dogs = {new_dog("bill"), 
                    new_dog("ben")} 

For me at least that is exactly the same as saying this "sequence" nonsense is just complete rubbish.

How many times do you think the average Euphoria program includes (say) misc.e, and how many programs do you think would stop
working if every include generated another separate instance, and what happens when filea includes fileb at the same time as
fileb including filea? It is not just daft and of no practical merit whatsoever, it is fundamentally impossible (to do "right"[2]).

[1] and within a routine, somehow magically sidestepping arbitrarily-nested routines, and of course with "variable namespaces".
I think everyone here knows I am no fan of namespaces, or rather using namespaces for anything other than their intended purpose
which is to resolve ambiguity, and certainly not in favour of deliberately introducing ambiguity so that namespaces are needed.
The Phix compiler, at least, applies common sense (ie locality) to automatically resolve conflicts whenever it can, which will almost
inevitably lead to some poor future maintainer making a wrong assumption about some code where namespaces have (accidentally)
been left off because they (presumably) did not carefully inspect every single line of code in the same way that the compiler has.
[2] where there are currently just over eight billion and counting completely different interpretations of "right".

PS footnote number one ran just a little off track there, didn't it?