Topic: Variable names with "#", "$" etc. Where does this come from?

[roy_043]

I have seen quite a few lisp examples where variable names contain a certain character to indicate the value type. For instance "#" indicates a variable that is used to store an integer, and "$" is used for string variables. Where does this come from?

[Kerry]

From people who didn't know how to use descriptive names and had no sense on aesthetics 

[Lee Mac]

Surely it's also bad practice as, if you decide to migrate to another API there's a good chance those characters aren't valid variable names...

[CAB]

I guess for purest it's like a stick in the eye.
I have seen some authors recommending this technique and thought at the time it was a good one but
gave it up in short order. For me the descriptive names worked better. The longer names ware me out typing
because I'm a four finger and two thumbs typist with the second worst spelling habits. So shorter names
get used now but I do try to keep them descriptive.

[Lee Mac]

I must admit, I do get lazy with variable names, for example, I used to represent an entity name with 'ent', over time this became 'en', and now its usually just 'e'. 

[Aerdvark]

I tend to be "not lazy" and use too long variablenames like "OldSystemVariables" or "NewInsertionPoint".
When making these two "OSV" or "NIP" I always end up with commenting it out:

(setq OSV ... ;_OldSystemVariables
NIP ... ;_NewInsertionPoint

Now the question rises why I use the underscore...?
I don't know, I sure did copy the idea from other lispers.

At one point I did concider the option to use $ for string variables but then there are so many variables that need a symbol.
$ for string, # for integer, % for real??, etc.

My 2 cents..

[mkweaver]

"$" to indicate a string variable comes from GW Basic - http://en.wikipedia.org/wiki/GW-BASIC

[Daniel J. Ellis]

I'm pretty sure that some programming languages required it (but can't remember any specifics).

Think it may have been recommended for BASIC, but that could just have been a dodgy memory!

dJE

[Aerdvark]

"$" to indicate a string variable comes from GW Basic - http://en.wikipedia.org/wiki/GW-BASIC

This reminded me of the Commodore 64 era....

[Keith™]

heh ... I typed a whole lotta keys while looking at that screen ... oh, the early 80s .. glad they are gone now

Now as far as the variables are concerned, I don't use things like $ or # in variable names, however, what I will do is use a prefix on my variables that indicate their type, for integers I might use a variable like intOffset, if for no other reason than to differentiate it from dblOffset ... I know I could make the names more descriptive, but I tire easily when I have really long descriptors.

Just wait till you get one of those files that use really wierd variable names .. like it is going to stop a knowledgable person from understanding the code. I've seen some real winners ..

Code: [Select]

(defun *wer$r (redewd) (princ redewd))

I mean seriously, who does that kind of crap anyway?

[Krushert]

Just wait till you get one of those files that use really wierd variable names .. like it is going to stop a knowledgable person from understanding the code. I've seen some real winners ..

Code: [Select]

(defun *wer$r (redewd) (princ redewd))

I mean seriously, who does that kind of crap anyway?

If ITwit/office manager wrote code, all the variables would names in Latin or my favorite, the names of Greek Gods.  When I hired on he had all the computers named names of the gods and demi-gods.  Yeah after my begging did not I had to threaten him with bodily harm.  :pissed:

[Matt__W]

I have seen quite a few lisp examples where variable names contain a certain character to indicate the value type. For instance "#" indicates a variable that is used to store an integer, and "$" is used for string variables. Where does this come from?

Most likely from encryption programs.  Convert.exe comes to mind.

[roy_043]

Thanks everybody for sharing your thoughts on this subject. I had the idea that this "convention" came from some other lisp dialect. Apparently not.

I disagree with Lee's comment about migrating. If you have to port a lisp routine to, let's say, some C-language, the variable names would be the least of your worries. Your biggest problem would be translating all the LISt Processing . But I know very, very little about C-languages.

As far as aesthetics are concerned, I only object to code that goes for a "kelvinated" look:

Code: [Select]

(defun c:XYZ ( /  A B C D E F G H)
....
I agree with CAB and Aerdvark and try to use short but clear names. In Aerdvark's example: OldSystemVariables or OSV, I would use oldSysVars.

<EDIT: typo>

[roy_043]

As an example, here is a link to some lisp code by Terry Miller who uses this convention.
http://web2.airmail.net/terrycad/AutoLISP-Code.htm

[Lee Mac]

I disagree with Lee's comment about migrating. If you have to port a lisp routine to, let's say, some C-language, the variable names would be the least of your worries.

Perhaps, but why add the problems...