regex poll 
Author Message
 regex poll

as ilya and i are having another war over MRE and versions of perl, i am
posting this poll. i want to know what regex features people use and how
often. i have listed all features i found in the desktop reference which
covers 5.005. if you see some that i have missed, add them. this is both
for s/// and m//.

this is for all levels of perl users. it is useful to know what both
newbies and gurus are using.

i want to see how many features are used and how often. i feel most
regexes use a core subset much more often than the more obscure features
(the 80/20 rule, 80% of regexes use only 20% of the features). this
subset (and more) is covered well by MRE and that is why i keep saying
it is useful regardless it being out of date regarding perl > 5.002. IMO
most of the regex features (by amount of usage) will be covered there.

answers: never almost-never rarely sometimes most-times often almost-always

dot .:                                  often
alternation |:                          sometimes
grouping ():                            almost-always
non-grabbing group (?:):                sometimes
back ref \1\2:                          never
group vars $1$2:                        often
special vars $`$&$':                        never
char class []:                          almost-always
class abbrev \d\D\s\S\w\W:              almost-always
quantifiers ?+*{}:                      almost-always
non-greedy ?:                           often
anchors ^$\A\Z\z:                       almost-always
look ahead/behind pos/neg (?=!<<!):       almost-never
embedded modifier (?misx):              never
embedded comments (?#):                 never
/x comments #:                          sometimes
/m modifier:                            sometimes
/s modifier:                            sometimes
/i modifier:                            sometimes
/x modifier:                            sometimes
/o modifier:                            rarely
/g modifier:                            sometimes
/c modifier:                            never
/e modifier:                            rarely
/ee modifier:                           rarely
pos function:                           rarely
alternate delimiters:                   sometimes
embedded code (?{}):                    never
conditional pattern (?()true|false):    never
anchored pattern (?>):                       never
regex object qr//:                      never
re pragmas:                             never

as a bonus what version(s) of perl do you use?

thanx,

uri

--
Uri Guttman                             Hacking Perl for Ironbridge Networks



Sat, 04 Aug 2001 03:00:00 GMT  
 regex poll

Quote:

>as ilya and i are having another war over MRE and versions of perl, i am
>posting this poll. i want to know what regex features people use and how
>often. i have listed all features i found in the desktop reference which
>covers 5.005. if you see some that i have missed, add them. this is both
>for s/// and m//.

Uri, as this poll is being conducted in an entirely unscientific manner,
the results will be useless for arguing for either position, whatever the
positions are.

--

%PGPKey = ('B76E72AD', [1024, '0824090B CE73CA10  1FF77F13 8180B6B6'])



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll
[A complimentary Cc of this posting was sent to Chris Nandor


Quote:

> >as ilya and i are having another war over MRE and versions of perl, i am
> >posting this poll. i want to know what regex features people use and how
> >often. i have listed all features i found in the desktop reference which
> >covers 5.005. if you see some that i have missed, add them. this is both
> >for s/// and m//.

> Uri, as this poll is being conducted in an entirely unscientific manner,
> the results will be useless for arguing for either position, whatever the
> positions are.

Definitely, but they might be entertaining nevertheless.

Ilya



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll



  >> Uri, as this poll is being conducted in an entirely unscientific
  >> manner, the results will be useless for arguing for either
  >> position, whatever the positions are.

  IZ> Definitely, but they might be entertaining nevertheless.

i think so too. i have about 10 responses. for some reason they are
coming to me and not the group even though i don't see a
followup-to-sender header. when i stop getting them (in a couple of
days) i will post a summary.

so far (by a quick visual scan) most of the fancy features are rarely
used.

and chris, if you don't like polls, don't reply. i never claimed
anything with this poll, i just asked for responses. people can make of
it what they wish.

uri

--
Uri Guttman  -----------------  SYStems ARCHitecture and Software Engineering
Perl Hacker for Hire  ----------------------  Perl, Internet, UNIX Consulting

The Best Search Engine on the Net -------------  http://www.northernlight.com



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll

Quote:

>and chris, if you don't like polls, don't reply. i never claimed
>anything with this poll, i just asked for responses. people can make of
>it what they wish.

I think you are speaking out of both sides of your mouth.  You
specifically stated that this poll sprung from a disagreement you and Ilya
are having about new regex features:

# as ilya and i are having another war over MRE and versions of perl, i am
# posting this poll.

The reasonable asumption is that you plan on using this data for
something, when that is completely unwarranted.  But now you say you are
not claiming anything with this poll.  This is a lie I've seen many times
in my years, in many forms: "this poll doesn't mean anything, but we spent
time on it and analyzed it anyway."  Why do it if it is a fact that the
results are meaningless?  I can only fathom that you are hoping that
meaningless facts will bolster your case against Ilya, whatever it is.
And that is Wrong, in the m{*filter*}sense, and it is Wrong, in the logical
sense.

I urge everyone to refuse to participate in such a thing.  There's no
knowledge to be gained by such a "poll".  It is entirely useless.  By
participating, you are legitimizing that which is de facto illegitimate.
If you have already sent your data to Uri, I urge you to ask him to not
include your data in his analysis.

Why am I doing this?  Uri contends this poll and its results have their
merits.  I contend they are meritless.  Uri intends, it is clear, to use
this data irresponsibly, to say that which cannot be said.  No matter what
his results are, they will be unindicative of any reality.

I have several goals in life.  One is to help stomp out the irresponsible
use of information.  This is a clear example of it.  I ask you all to help
by not participating.

I've said my peace, I'm done.

--

%PGPKey = ('B76E72AD', [1024, '0824090B CE73CA10  1FF77F13 8180B6B6'])



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll
[A complimentary Cc of this posting was sent to Chris Nandor


Quote:
> Why am I doing this?  Uri contends this poll and its results have their
> merits.  I contend they are meritless.

This would not make them so.  Say, if there are people who use (?{})
already, I may want to contact them to discuss their uses so that I
may improve/optimize it more.  

With Hugo's patch for 5.00555 (?{}) has a pretty good chance to become
quite handy.  But in some restricted forms it could be used even before.

Ilya



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll
(Newsgroups trimmed to .moderated only)


Quote:

> I urge everyone to refuse to participate in such a thing.  There's no
> knowledge to be gained by such a "poll".  It is entirely useless.  By
> participating, you are legitimizing that which is de facto illegitimate.

Whoops... well, I responded before I realized that the poll was
controversial, but it appears my response has been canceled by someone --
one of the moderators, I assume? -- in any case.

In Uri's defense, I can say that I only responded because (a) it was easy,
(b) I'm kind of curious myself as to how others use regexes (especially to
see if I'm missing anything cool that everybody's using except me), and
(c) I don't understand the deeper issues that Uri and Ilya are debating,
and so I wasn't worried about submitting biased information.

It seemed like a simple enough request.  If Uri had in mind the end goal
of using information from this informal poll in an argument against Ilya
(or whomever), I am sure many people would point out the unscientific
nature of his data, no?

--
Steve Linberg, Systems Programmer &c.
National Center on {*filter*} Literacy, University of Pennsylvania

WWW: < http://www.*-*-*.com/ >



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll
I want to know if there is any connection between this poll being
a disinformation campaign and the fact that my "survey" answers
were rejected. I ask this for one because a near-identical post has
appeared on the group, but mainly because I want to know how a Perl
programmer's post about his programming habits is inappropriate. It
doesn't sound like a borderline case on which different moderators
might disagree.  

I will certainly gain something from seeing certain people's answers,
namely people about whose personalities and concerns I know something
by reading the group. If I find someone who is exactly like me but
uses one feature far more than I do, I know what to play with next.
If I see someone in the web business not using a "useful" feature
at all, I could ask them how they solve certain problems... and
maybe end up learning another way.

And perhaps the poll was intended to sway The Mighty Zakharevich,
but the possibility of that happening is laughably small. It's just
not the kind of energy which gets rechannelled by a voice vote,
that's the valuable thing about it...



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll

Quote:
> [...]
> I have several goals in life.  One is to help stomp out the irresponsible
> use of information.  This is a clear example of it.

Well, to the purpose of Uri's summarizing things for the group, I
think this is simply an example of casual (unscientific) collection
and dissemination of some interesting information.  If Uri
subsequently tries to argue a point with Ilya based on this
information, then that's arguably irresponsible, but that's really
Uri's and Ilya's business.

Quote:
> I ask you all to help by not participating.

I fully intend to e-mail Uri my responses, and I'm looking forward to
seeing the results.  (I hope Uri posts them, even if he only gets 20
responses.)  I think seeing even 5 responses could be useful
information.  Everyone has to interpret it their own way, of course.
For example, I hope the data can be presented in raw form as well, so
readers have the option to normalize the data sets before calculating
averages or other statistical features.

--Todd

p.s.  By reading Uri's list, I just learned of the existence of a couple
recentish constructs that I didn't know existed.



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll

Quote:
> I want to know if there is any connection between this poll being a
> disinformation campaign and the fact that my "survey" answers were
> rejected.

No.  In retrospect, I probably shouldn't have approved Chris's message
about polls either; that was my call, and I'll take responsibility for
it.  The wheres and whys of polling really aren't on topic in a Perl
newsgroup.

The reason for not posting the survey answers is that Uri is taking the
survey and presumably will collect responses and summarize; that's the
standard way on Usenet that such things have always been done for decades.
Posting all of the individual responses *plus* the final summary just
results in lots more traffic without much additional signal.

If people have general *discussion* of regex features, of course that's on
topic.  It's specifically the poll responses that the moderation team
collectively discussed and decided weren't of sufficient general interest
individually in light of an eventual summary.

Quote:
> I ask this for one because a near-identical post has appeared on the
> group,

Which was a mistake; apologies for that.

Quote:
> but mainly because I want to know how a Perl programmer's post about his
> programming habits is inappropriate.

It's not.  It's the general context of the survey and survey responses,
and the fact that survey responses in and of themselves aren't discussion,
just lists of regex features you use.

Quote:
> I will certainly gain something from seeing certain people's answers,
> namely people about whose personalities and concerns I know something by
> reading the group.

I think it would be great if Uri would put that information in the
summary.

Quote:
> And perhaps the poll was intended to sway The Mighty Zakharevich, but
> the possibility of that happening is laughably small. It's just not the
> kind of energy which gets rechannelled by a voice vote, that's the
> valuable thing about it...

*grin*

For the record, these are all my personal opinions and personal
understandings; I'm not speaking for anyone other than myself.

--
#!/usr/bin/perl -- Russ Allbery, Just Another Perl Hacker





Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll



  >> Why am I doing this?  Uri contends this poll and its results have their
  >> merits.  I contend they are meritless.

  IZ> This would not make them so.  Say, if there are people who use (?{})
  IZ> already, I may want to contact them to discuss their uses so that I
  IZ> may improve/optimize it more.  

but since you plonked me, you won't be able to read the results when i
post them!

:-)

uri

--
Uri Guttman                             Hacking Perl for Ironbridge Networks



Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll

Tushar Samant writesL

Quote:
> I want to know if there is any connection between this poll being
> a disinformation campaign and the fact that my "survey" answers
> were rejected. I ask this for one because a near-identical post has
> appeared on the group,

That post was approved by mistake.  It's been cancelled, and we asked
the author to forward a copy directly to Uri.  On behalf of the panel,
I apologize for any confusion that this caused.


Sun, 05 Aug 2001 03:00:00 GMT  
 regex poll

Quote:

>The reason for not posting the survey answers is that Uri is taking the
>survey and presumably will collect responses and summarize; that's the
>standard way on Usenet that such things have always been done for decades.

Sorry, I didn't think of that angle.


Tue, 07 Aug 2001 03:00:00 GMT  
 regex poll

well, the responses finally trickled away so i got off my duff and
generated a complex statistical analysis. after poring over the data and
spending hours of cpu time crunching the numbers, i have come to the
irrefutable conclusion that perl hackers overwhelmingly prefer Crelm
toothpaste with the miracle ingredient Frauduline! (hi chris :-)

we got 37 replies (including myself) and i did have to spend some time
cleaning up the replies so i could parse them with a simple script. did
you really think i wanted to crunch the text by hand? why did some of
you answer with your own formats, answers, in line comments, etc? anyhow
i finished it and here are the results. if your name is not in the list
at the end i may have misplaced your reply or i never got it. send it
again and i will add it to the results. if i really get my act in gear i
may webify the poll and results, but i doubt that will happen soon.

i won't make a major comment on the results (other than the obvious
toothpaste preference), but it does seem like most regexes use a fairly
well defined subset of the available features. my main point is that in
the real world you only need the basic regex stuff crunch and parse
text. many of the newer features either aren't need in most cases or are
not well understood or are not even available.

anyway, make of this info as you will. it is not scientific, accurate,
legal, socially responsible, or worth any money. but it does contain
information which is all number 6 had.

uri

here is a simple histogram of versions i extracted from the replies.

5.001_          1
5.002           6
5.003           4
5.004           5
5.004_          2
5.004_01        1
5.004_02        1
5.004_03        3
5.004_04        16
5.005           6
5.005_01        1
5.005_02        17
5.005_03        2
5.005_53        1
5.005_54        2
5.005_55        1

these are the answer keys for the counts below them

                   never
                       almost-never
                           rarely
                               sometimes
                                   most-times
                                       often
                                           almost-always

dot .
                   0   0   0   1   3  18  16

alternation |
                   0   1   5  21   1  10   0

grouping ()
                   0   0   0   2   2  11  23

non-grabbing group (?:)
                   8   5   6  14   0   5   0

back ref \1\2
                  15   9   5   7   1   1   0

group vars $1$2
                   0   0   0   7   4  22   5

special vars $`$&$'
                  22   8   3   3   0   2   0

char class []
                   0   0   0   5   1  13  19

class abbrev \d\D\s\S\w\W
                   0   0   0   2   1   4  31

quantifiers ?+*{}
                   0   0   0   0   0   1  37

non-greedy ?
                   2   3   4  12   0  16   1

anchors ^$\A\Z\z
                   1   0   0   0   0  10  27

look ahead/behind pos/neg (?=!<<!)
                  15  15   6   2   0   0   0

embedded modifier (?misx)
                  29   6   2   1   0   0   0

embedded comments (?#)
                  31   4   1   1   0   1   0

/x comments #
                  11   7   4  13   0   2   1

/m modifier
                   7   3  12  16   0   0   0

/s modifier
                   7   2  12  16   0   1   0

/i modifier
                   1   1   6  14   4  12   0

/x modifier
                   4   6   5  17   2   4   0

/o modifier
                   6  11  16   2   1   2   0

/g modifier
                   0   1   1  21   6   9   0

/c modifier
                  33   4   0   1   0   0   0

/e modifier
                   9   5  14   8   0   2   0

/ee modifier
                  23   5  10   0   0   0   0

pos function
                  19   8   8   3   0   0   0

alternate delimiters
                   1   1   2  19   4   9   2

embedded code (?{})
                  34   2   2   0   0   0   0

conditional pattern (?()true|false)
                  35   1   2   0   0   0   0

anchored pattern (?>)
                  35   1   1   0   0   0   1

regex object qr//
                  28   2   4   3   0   1   0

re pragmas
                  36   2   0   0   0   0   0






































--
Uri Guttman  -----------------  SYStems ARCHitecture and Software Engineering
Perl Hacker for Hire  ----------------------  Perl, Internet, UNIX Consulting

The Best Search Engine on the Net -------------  http://www.northernlight.com



Wed, 15 Aug 2001 03:00:00 GMT  
 
 [ 19 post ]  Go to page: [1] [2]

 Relevant Pages 

1. regex poll

2. ADMIN: Uri's regex poll

3. Sybperl poll

4. Free CGI Poll - check it out

5. Free CGI Poll - check it out

6. Text Version Poll

7. STRAW POLL: RFD: comp.lang.perl reorg

8. Poll scripts how ?

9. Polling exit values?

10. Poll: Please Take 2 Minutes to Answer...

11. adding admin feature to poll script

12. polling TCP server on Win32 (with Tk)

 

 
Powered by phpBB® Forum Software