Re: $:attr vs $.:attr
On Sun, 2005-05-15 at 19:18 +0200, Juerd wrote: Now: Declaration ExplicitImplicit $_ $?SELF has $.var | $obj.var \ .var \ ./var \ has $:var | $obj.:var \ .:var \ ./:var \ Consistent: has $.var \ $obj.var \ .var \ ./var \ has $.:var \$obj.:var \ .:var \ ./:var \ See it yet? It's about consistency in the whole scheme, not the clarity of a single element. I'm not sure I see that you changed anything other than changing Old way to Now, Your proposal to Consistent, adding separator characters between the columns, and removing the word Implicit from the last column. I'm still not seeing the problem. Counting down and then right, I read those: $. means attribute; $: means private attribute; $x. means accessor method; $x.: means private accessor method; and so on... I think you're confusing the method glyph . and the attribute glyph, also . in the first and second columns. Consistency would be to make accessor methods use both the attribute and accessor glyphs .., but that would cause at least 4 problems that I can think of, so . is a nice simple thing to do. In the case of private accessor methods, most of those reasons go away, so .: needn't be (and in fact can't be) shortened. All of the details aside, I simply don't see newbies getting all that confused. For starters, there's no particular reason that they would have to use private attributes all that often, and if they did, they don't really EVER have to use accessor methods on them, though style might nudge them in that direction, there's no real impact on scalability the way there is with regular accessors (because every use of the private accessor is local to the class definition). Now, the one place I can see that you should ALWAYS use the private attribute accessors is in roles. Here, it would be essential, as you don't know what the composing class might want to do to that method, so the only place that I see this being a big concern is in roles that have private attributes. Certainly an edge case newbies won't contend with all that often, yes? Note that it's not *implicit* $?SELF. ./ is a prefix operator that calls a method on $?SELF That's picking nits. I know that, but was referring to the implicit use of $?SELF regardless of the syntax one would have to employ in order to use it explicitly.
Re: $:attr vs $.:attr
Aaron Sherman skribis 2005-05-16 5:54 (-0400): On Sun, 2005-05-15 at 19:18 +0200, Juerd wrote: Now: Declaration ExplicitImplicit $_ $?SELF has $.var | $obj.var \ .var \ ./var \ has $:var | $obj.:var \ .:var \ ./:var \ Consistent: has $.var \ $obj.var \ .var \ ./var \ has $.:var \$obj.:var \ .:var \ ./:var \ See it yet? It's about consistency in the whole scheme, not the clarity of a single element. I'm not sure I see that you changed anything other than changing Old way to Now, Your proposal to Consistent, adding separator characters between the columns, and removing the word Implicit from the last column. Okay, let's try it differently, then: Now: Declaration ExplicitImplicit $_ $?SELF .| .\ .\ . \ :| .:\ .:\ . :\ Consistent: .\ .\ .\ . \ .:\.:\ .:\ . :\ I think you're confusing the method glyph . and the attribute glyph, also . in the first and second columns. It's the same thing. $.foo installs an accessor *method*. $object.foo calls that method, it doesn't access $object's $.foo directly. Note that it's not *implicit* $?SELF. ./ is a prefix operator that calls a method on $?SELF That's picking nits. In the process of designing something, every single nit must be picked. Don't take it personally. Juerd -- http://convolution.nl/maak_juerd_blij.html http://convolution.nl/make_juerd_happy.html http://convolution.nl/gajigu_juerd_n.html
Re: $:attr vs $.:attr
On Mon, 2005-05-16 at 12:23 +0200, Juerd wrote: Aaron Sherman skribis 2005-05-16 5:54 (-0400): I'm not sure I see that you changed anything [...] Okay, let's try it differently, then: [...something that looks like braille...] And now, you've s/[\$\w]+//g; what point are you making, Juerd? I think I've demonstrated that I get it. I just don't agree with your thesis that that one character is a) a problem or b) inconsistent. I think you're confusing the method glyph . and the attribute glyph, also . in the first and second columns. It's the same thing. Well, there's the problem. If you see those as the same symbol, then of course this is deeply confusing. After all, $.foo doesn't behave anything like a method, and yet it uses the method symbol. Try looking at those two symbols as TWO SYMBOLS, and I think your life gets easier. $.foo -- Attribute glyph . used after scalar variable glyph $. $obj.foo -- Method glyph . used after scalar variable. If you de-couple those in your mind, then there's $.foo installs an accessor *method*. Well, it can. It won't if there's already a method there, of course, and that's an important distinction. There's no fundamental tie between the two. The auto-accessor is just a convenience that is placed there for you if you didn't go out of your way to supply one. It's not part of the attribute, and so we should not name the attribute in order to match the naming of the method. If anything, we should do the reverse, but I'm happy with the way it is. $object.foo calls that method, it doesn't access $object's $.foo directly. Yes, that's correct. That's because you're using the method-invocation glyph. If you had used the attribute glyph ($.foo), then you would get direct access. Are you telling me this because you didn't think I knew? It certainly doesn't seem to further the point Note that it's not *implicit* $?SELF. ./ is a prefix operator that calls a method on $?SELF That's picking nits. In the process of designing something, every single nit must be picked. Don't take it personally. You removed the context supplied by my next sentence. Your nit to which I was referring was NOT a design comment, and I still hold that my original usage of implicit is correct.
Re: $:attr vs $.:attr
Now: Declaration ExplicitImplicit $_ $?SELF has $.var | $obj.var \ .var \ ./var \ has $:var | $obj.:var \ .:var \ ./:var \ Consistent: has $.var \ $obj.var \ .var \ ./var \ has $.:var \$obj.:var \ .:var \ ./:var \ See it yet? It's about consistency in the whole scheme, not the clarity of a single element. (I wouldn't mind if $:var was a shortcut syntax for $.:var, although we will regret this when we think of an even better purpose for the colon as a sigil.) Note that it's not *implicit* $?SELF. ./ is a prefix operator that calls a method on $?SELF, not an infix operator that when prefixly used defaults to something. ./ is not like .+ and friends. It cannot be used infix, it does not default to anything. Read the two characters as one. Juerd -- http://convolution.nl/maak_juerd_blij.html http://convolution.nl/make_juerd_happy.html http://convolution.nl/gajigu_juerd_n.html
