Yes, the worst ones.

But love and hate are emotions. Let’s talk facts.

Re 1.: “foo” is just a string (which could be the name of a variable. “$foo” is the value of the variable foo, as long known from shells. The difference between a name and a value ($) gets clearly distinguished. (In Lisp you’d need “special forms” to avoid evaluating everything…)

Re 2.: Loop and if bodies, as well as procedures, are compiled to bytecode at first use (since Tcl 8.0). On later calls, only the bytecode is executed again.

Re 3.: I’m often fixing bugs with a quick glance at the backtrace. They look clearer to me than Python’s.

Re 7.: [expr] was introduced as a compromise so expressions could be written in infix notation ($a + 1). It is easy to add prefix notation that’s more in the Lisp/Tcl spirit, so one can write [+ $a 1] if one wishes.

Re 8.: Conditions in [if], [while] and [for] are expressions, so it’s a good move to give them an implicit [expr].

Re 9.: Every value has a string representation, and possibly one of another type, in parallel. Powerful data structures include lists (that can be nested to make matrices or trees), hash tables and dictionaries.

In short: Tcl combines concepts from Lisp, C, and shells. And often it can be even more powerful than any of these.

Tcl syntax is, in fact, very consistent. It’s just that consistency is
overrated. You know what they say about a foolish consistency…. What
you’re actually saying is that Tcl doesn’t match the way you like to
think about variables and program statements. That’s as good a reason
as any not to like Tcl.

Oh my god.

TWELVE YEARS.

– Dave

Pick Basic (the programming language for Pick-style databases) does a very nifty string-as-data-structure trick.

X = “A;B;C”
CONVERT “;” TO @FM IN X
CRT X<2>

This turns the semi-colons in the string into Field Marks (@FM), then uses the field extract operator (“<>”) to pull out a specific field. It’s even more elegant

It also does a very neat integer-as-string trick too:

FOR X = 1 TO 1000
IF X[2] = “00″ THEN
CRT “*”:
END
NEXT

Here I’m using X as an integer in my FOR loop, then later on I’m using it as a string by using the substring operator (“[]“). This prints a * when the last two characters of X are “00″, which is on every 100th pass through the loop.

Plus it has other features, such as compiling (to bytecode, no less), printing out exactly where an error happened (but no backtrace), and a few others.

And it dates back to sometime in the late 60s, IIRC. It has evolved in the meantime, of course, so things like bytecode might have come later.

– Dave