Call for signatories for J2

eltronic at juno.com eltronic at juno.com
Thu Aug 26 02:11:01 CEST 2004


please count as for the J2 proposal of yesterday.
please list as anon, email withheld.
or better yet email mangled, as this and 
other lists promise to mangle but they never do.
here comes another 100 spams and virus as a result of this post.


H J2 A
H (|deco | deco instead of @deco)
J2 (word: suite where word is decor, using, other)
    include time for objections to any word picked.
    note, many are on vacation or only sporadically
    even read c.l.py and less frequently py-dev.
    a python-announce post woulden't be a bad idea.
A (@deco, @ deco)

H checked in implementation already exists
  that can be modified with minimal effort.
  solves some problems for Leo,
 but long range, since Leo edits multiple languages
 and Java, assembler, forth, perl others' use @
 Leo will continue working somehow.
 by avoiding existing and user added @directives.
 by Leo allowing you to define the @ to something else.
 by @verbatim. 
 any change is going to have consequences.
  not sure about other tools, Ipython, empy.
  learning new behavior for users, all the
  supporting doc amendments, good search for
  prior help tips out the window. in addition
  to the trivial by comparison code changes.

  @@ will interfere with diff listings.
  texi format uses @directives? has anyone checked?
  is it really the position of the BDFL that
  python must use @ at any cost? when | works.
  was surprise a necessary part of the strategy?

  folding editors and code import no worse off
  than how they now handle f = staticmethod(f).
  since they probably don't now alert the user
  there is something special about this def, 
  they needed to be fixed anyway. no doubt 
  all tools will need upgrade whichever. we could 
  single out a few that aren't up to py 2.2 yet.
  that don't differentiate old from new classes.

  | generates syntax error? or would it be unary?
  why is it a problem that | looks like l or 1 
  on some fonts? and might '?' be another alternate char?
  but I will not try to find the answer if its already
  posted, imagine trying to lookup python related 
  pages including ? or @, 
  some searchers commonly ignore punctuation.
  are decorators destined to be forever obscured?
  these are not inconsequential questions!
  in the few weeks since wider public exposure
  dozens of such questions exposed. even the omnipotent
  can't keep them all on the same page at the same time.
  more than a few interested parties just tuned out early on.
  
J2 has tacit community approval and implementation.
 good job on the pep, the wiki, the vote and proposal.
  since vote wasn't secret, probably many just
   decided to minimize bandwidth and held off posting.
   the one lopsided poll, evidence of nothing.
   the correct result somehow emerged.
 favorable words, 
      using, deco, decor, trans, alter,
      declare, combine might already be used as var names.
      implement, may be better left for something else.
      with & as, overloaded, but actually not too bad.

@deco is a syntax that will be tolerated, but,
how can so many negatives be a positive?
presumable we will be able to try/except 
if no def found, I have seen no mention of this.
will there be a specific Exception?

is there unicode or other popular name 
for @ besides NULL or at?

the real shocker in all this is how few dev people
are aware of how @deco affects Leo & IPython.
guess the word is not getting out how much they 
are an improvement on the basic python intrepreter.
oh, You dont realize Leo executes python? much pity.
it is completely scriptable in nearly every way.
but I guess you can cobble something together similar
in emacs or vim, not to put too fine a point on it.

decor:
    staticmethod
def whatever():
    pass
anything more complicated than that I wince.
there are people with a use case who say they 
need it, we all really should get out of the way.
I had occasion to revisit a few methods I
moved out of a class because they didn't need
to be methods. I ran some quick timing tests
and found staticmethod accessed from other 
class members appear to be faster so I made
them staticmethod. I know its heresy to
support a change like this based on speed,
based on timing of toy pared down examples,
but maybe the decorator would be seen as
more directly affecting people if they realize
how they can use the thing. the docs, 
none of the available lit even briefly
touch on this aspect. or maybe I gloss over it.
too few comparing and contrasting examples.
need more like @memoize in a few of the posts.
python is deceptively simple, you can
go a long time with just the basic simple stuff.
another case in point, it took a very long time
to feel comfortable with slices and list 
comprehensions. in spite of 20 years of c.
now, I see their application directly as simple.
take the advice of a few early posters in this 
by far more archive building issue than 308 
and try decorators out to see them in action.
or the also good advice, to make your case, then 
stand back as the creators of Leo and IPy have.

ask me again in 5 minutes I might say
no change is justified. stop the insanity. but,
the case for waiting is a fruitless non argument.

recommend save pep numbers ending in 8
for only the most contentious issues.
is 0332 being saved for y2032 issues?
0338 anyone up for macros? do we get
@deco then ?macro soon after?
do we get |deco or ?deco then @macro
I sense a fork forming in a few minds.
I am amused that $ slipped in as an % 
interpolation replacement under the radar.

________________________________________________________________
The best thing to hit the Internet in years - Juno SpeedBand!
Surf the Web up to FIVE TIMES FASTER!
Only $14.95/ month - visit www.juno.com to sign up today!



More information about the Python-list mailing list