Notes

8 Responses to “Rmvng ll vwls”

1. Weronika Says:
   January 11th, 2009 at 9:46 pm

   Eh, it’s reasonably readable. Goes slower than normally, but I can get the general idea in I’d say no more than 5x the time for normal text, and I can figure out all the words if I think about them for a bit.

2. irving Says:
   January 11th, 2009 at 9:58 pm

   My ctirorien for rbeaalde was taht I culod raed it at rlohguy the smae speed. Three etxsis srnispurig ttaooanmfnsrirs taht do have that ptpreroy, so I was holndig it to a hieghr sdatrnad.

3. Weronika Says:
   January 11th, 2009 at 10:02 pm

   There do, really? What are they? That seems highly unlikely.

   For me the two transformations here are on a similar readability level - how about you? I wonder if it’s a native language issue.

4. irving Says:
   January 11th, 2009 at 10:12 pm

   That’s quite possible, since Polish is a much stricter language spelling-wise. I think rolmnady shuilnffg all but the first and lsat wrods is much easier to read than rmvng ll th vwls in large part because it mangles short words less, but I can’t make a very scientific comparison if I’m trying to read scrambled text that I wrote.

5. Weronika Says:
   January 11th, 2009 at 10:31 pm

   It could be either Polish, or just knowing more languages and being used to guessing ones I don’t actually know. The scrambled words frequently look like potentially real words in some language, which slows me down. The no-vowel ones are clearly not real words.

   Also - are you doing this by hand, or do you have a program? If you have one, you could send it to me and I could generate you some scrambled text.

6. irving Says:
   January 11th, 2009 at 10:35 pm

   Here’s a vowel removing script:

   #!/usr/bin/python
   
   import re
   import fileinput
   
   pattern = re.compile(r'[aeiouAEIOU]')
   for line in fileinput.input():
       print pattern.sub('',line),
   

   and a middle scrambling one:

   #!/usr/bin/python
   
   import re
   import random
   import fileinput
   
   def scramble(s):
       mid=list(s[1:-1])
       random.shuffle(mid)
       return s[0]+''.join(mid)+s[-1]
   
   for line in fileinput.input():
       line=re.sub(r'\b(\w\w\w\w+)\b',lambda x:scramble(x.group(1)),line)
       print line,
   

7. Weronika Says:
   January 12th, 2009 at 12:08 am

   Using different text for the two transformations, two passages each.

   Th rllng f sx-sdd dc n ntrl cndtns my b sd t prdc rndm rslts n tht n cnnt cmpt bfr rll wht dgt wll b lndd n, bt th prbblty f lndng n ny f th sx rllbl dgts cn b clcltd bcs f th fnt crdnlty f th st f pssbl tcms.

   W wnt t Flrd, whch ws… trplk. rlly, rlly dn’t lk trvlng n rd-y t spnd svrl dys n frcd cmpnnshp wth grp f ppl. Tht’s th lst tm lt smn ls hndl pln tckts, spclly smn wh cn pprntly gt fll nght’s slp n tw 3-hr pln flghts.

   A crtneal qesiuton of comutepr sceince is to aesrdds the lmtiis of coipmtnug decevis. One aoracpph to arednidssg tihs qitseoun is ursedtidnanng the pmrbeols we can use cpoumrtes to svloe. Meodrn ctpnuiomg deevics oetfn seem to pesssos intiifne ctaaicpy for caclatlioun, and it’s esay to igimane that, gevin egnouh tmie, we mhigt use cmroputes to svole any plerobm.

   Srroy abuot the bnlak eiaml earleir. My mial’s been gnoig sort of funky in resopsne to commdans eevr scnie it strtaed sespuntolnoay qtniuitg ydtareesy bucasee I eexeecdd the szie limit. It then wdolun’t let me deltee ealims, ehetir, buesace it didn’t hvae the memroy to do that. So I tnihk I snet the bnalk eimal as an aciecndt.

8. Weronika Says:
   January 12th, 2009 at 12:15 am

   Hmm, interesting. For me no-vowels is easier in short passages, but middle-scramble quickly gets easier in longer ones, once I know the context.

   Also, reading these has strange effects on reading normal text.

Leave a Reply

Name (required)

Mail (will not be published) (required)

Website

Notify me of followup comments via e-mail