Science a GoGo's Home Page Forums General Discussion General Science Discussion Forum Try to read this

Are you rnaedig the news?
7 July 2006

If you tnhik you're reandig the news, be warned taht this stroy -- and any ohter on the web -- wlil be berlay read by anynoe 36 hours aeftr it was frist poetsd. Taht's the megasse form a taem of stcttiaisal pytsichiss who hvae asnlaeyd how pelope ascecs iftoaorminn olinne. Aeblrt-L?zsl? B?sabari of the Uvirisntey of Notre Dmae in the US and cueaeolgls in Hruangy have culelaatcd that the neubmr of pelpoe who read nwes sreiots on the web dcyeas wtih tmie in a power law, and not etloxplnaeniy as colnmomy thhugot. Most news bcemeos old hat wihtin a day and a hlaf of bineg pstoed -- a fninidg that could hlep wetbise deriegsns or pelpoe tyinrg to uasdnertnd how imtifonaron gtes trfrnreased in bocaloigil cells and soical nrekowts (Phys. Rev. E 73 066132).

Psitiyshcs like Bba?asri are ierntetsed in stynduig the Wrold Wide Web baecuse it is an exlpmae of a "cmoelpx ntroewk", wtih a topolgoy that chgenas as new douetcmns and likns are culnanloity aeddd. His team preiucts a tcayipl nwes web stie as a sereis of ciurclar bblos, or "nodes", ecah of which cesrnpdroos to an idnavdiuil nwes sorty, with a line jinniog ecah ndoe if the two seotris are cconeentd by a hypelnirk (see frguie). The aera of ecah blob is paprnotoriol to the lorhatigm of the nuembr of vsiits to ecah domcuent.

Tiehr model reveals taht a tpcayil nwes stie has a raevlltiey sabtle "sektloen" -- crnrpesoodnig to the ovrlael oantaozrgiin of the stie -- aolng wtih nodes (that is, aatucl seotirs) that are only toiapermrly lienkd to the main srcturute brfeoe bneig deetled from the stie or not lnekid any mroe. In this sense, the nretwok rmeebesls a biaocoligl clel's rgaltreuoy newtork, woshe "wrinig" can cahnge rpalidy dnriug a cell cycle. It is aslo a bit lkie socail nrokewts: we ecah hvae a raiveletly sabtle core noewtrk of feirnds and ainecqancutas but the nbeumr of poplee we ianrtcet wtih can vray diatlslrcay from one day to the next.

To get a fellur undrdsneniatg of such newotkrs, Bbas?rai and cleoaluegs deiecdd to sutdy the vsniitig prteatns on a popualr Huainrgan nwes and eenatnmeintrt patorl (oirgo.hu). Tkahns to aauoalmctitly anssgeid "ciookes", the stiscnties wree albe to rercuctnost the boswring hitrsoy of aobut 250,000 vsritois to the stie oevr the cuorse of a mnoth.

The rcasrehrees fonud that the dmenctuos bnenilogg to the skoleten of the wbeitse reevice an axtlmaipopery csnoatnt setarm of virotiss, wichh means that the cmvuituale nbmuer of vsrtoiis aseicscng thsee dtcneomus iacsrnees lriaenly in time. In cnsotart, the nwes decmnouts rvcieee the msot hits diretcly aeftr their reeslae, and derecase with time. Tuhs, the cuumvaltie nrebmus of vitiss hree raech siauottran aetfr jsut a few dyas.

Braasbai's taem cetcualald the "hlaf-life" of a nwes doencumt, wchih cdsoorrnpes to the prioed in wchih half of all vsirtios taht enalevulty aseccs it have vsiietd. The rreshcarees fnoud taht the oearvll half-lfie ditbutsiiorn fwlools a poewr law, wihch iaiedtncs taht most nwes ietms have a very srhot lftmieie, aluhtgoh a few cuoitnne to be asececsd wlel boyend this peroid. The aagvree half-lfie of a news item is just 36 huors, or one and a half days afetr it is reasleed. Wilhe this is short, it is lgeonr tahn pdeierctd by spmile etaoepixnnl meodls, wihch assume that web page biswrong is less rdnaom tahn it aclltuay is.

The shrot life of a nwes item -- coenmbid wtih roadnm vnstiiig prtaents of reerads -- iipmels taht poelpe cloud miss a siifningcat ftioacrn of nwes by not vinsiitg the potarl wehn a new dmucnoet is fisrt dyaespild, wihch is why pulrihbses like to pdrovie e-mial news aletrs. The rtlsues aslo show that peolpe read a putalaicrr web page not jsut bacseue it looks itsetninreg but buacsee it can be aseccsed eailsy.

Ahgotulh the avaerge hlaf-life vaeirs for dienfreft types of sites, the decay lwas iiefeitndd are llieky to be giernec buesace they do not dneped on cennott, but are mnlay dmtenreied by a user's viniitsg and bnorsiwg prtatens.

"Such qanivuatttie approhecas to olnnie mdeia not olny oeffr a beettr uanetiddnrnsg of inmriatfoon aeccss, but culod have iatnprmot croemimacl ailcinatppos as wlel ? from betetr partol dsgien to usatdndninerg irionmotafn diosfifun, folw, and miktnaerg in the onnile wolrd," say the rerseahcers.

Y?

This gibberish should be deleted. And I think I will do just that, after 24 hours. The language of this forum is English, not gibberish. I find this makes no sense at all.

Amaranth
Moderator

Amaranth, I think that's the point!

does not matter. deliberate gibberish is not right. doing it once would be ok, but there have been several post in several threads about this. its not worth reading.

actually taking the time to read it, there is a little content worth gleaning

not having encountered it before, i find it an interesting form of intentionally dyslexic writing (seems like someone worked hard to scramble it that bad, but make it so that it can readily be unscrambled by the native english speaking mind)

interesting how the non-dyslexic mind automatically sorts it out making it legible

hmm...reads kind of like some of dano's posts, but a lot more sensible :-)

ok, stop that...baaaad daawgie

I saw this first at NGC channel (inbetween two programs they sometimes show adds about their own programs). If you randomly permute the letters of words but keep the first and last letter of each word in its original place (E.g. height --> hgeiht), then you will be able to read most of the words. And the faster you try to read the text the easier it is. If you switch to ''fast reading mode'' the brain pays less attention to the order of the letters in a word making it easier to read the scrambled text.

On NGC channel they showed a few sentences of scrambled text, but I decided to try it out for myself, so yesterday I wrote my own program to see if it really works as advertised. I later found out that there is a Java program available for scrambling text:

http://www.ramdhanyk.com/movabletype/archives/thoughtprocess/001378.html

It's not so difficult using a scrambling program . If you have a pascal compiler (you can download one from here ) you can compile this text scrambler program:


PROGRAM SCRAM;
VAR CH: CHAR;
TEST:BOOLEAN;
INF,OUTF: TEXT;
K,TEL: BYTE;
ARPERM: ARRAY OF BYTE;
ARCH: ARRAY OF CHAR;
FUNCTION TESTLET(X:CHAR):BOOLEAN;
BEGIN
CASE X OF
#65..#90,#97..#122,#128..#151,#153,#154,#159..#165,#181..#183,
#198,#199,#210..#212,#214..#216,#224..#229,#233..#237: TESTLET:=TRUE
ELSE
TESTLET:=FALSE;
END;
END;
PROCEDURE PRM(N:BYTE;VAR AR:ARRAY OF BYTE);
VAR INTM,K,R:BYTE;
BEGIN
FOR K:=0 TO N-1 DO
AR[K]:=K;
FOR K:=0 TO N -2 DO
BEGIN
R:=RANDOM(N-K)+K;
IF R>K THEN
BEGIN
INTM:=AR[K];
AR[K]:=AR[R];
AR[R]:=INTM;
END;
END;
END;
BEGIN
RANDOMIZE;
ASSIGN(INF,'IN.TXT');
ASSIGN(OUTF,'OUT.TXT');
REWRITE(OUTF);
RESET(INF);
REPEAT
READ(INF,CH);
IF TESTLET(CH) AND NOT(EOF(INF))
THEN
BEGIN
TEL:=0;
REPEAT
TEL:=TEL+1;
SETLENGTH(ARCH,TEL);
ARCH[TEL-1]:=CH;
READ(INF,CH);
TEST:=TESTLET(CH);
UNTIL NOT(TEST) OR EOF(INF);
IF TEST THEN BEGIN
TEL:=TEL+1;
ARCH[TEL-1]:=CH;
END;
IF TEL>3 THEN BEGIN
SETLENGTH(ARPERM,TEL-2);
PRM(TEL-2,ARPERM);
WRITE(OUTF,ARCH[0]);
FOR K:=0 TO TEL-3 DO WRITE(OUTF,ARCH[ARPERM[K]+1]);
WRITE(OUTF,ARCH[TEL-1]);
END
ELSE FOR K:=0 TO TEL-1 DO WRITE(OUTF,ARCH[K]);
IF NOT(TEST) THEN WRITE(OUTF,CH);
END
ELSE WRITE(OUTF,CH);
UNTIL EOF(INF);
CLOSE(INF);
CLOSE(OUTF);

It helps to pctriace your fast rneiadg sillks.

No, tihs is fun. Many plpeoe wlil wnat to try it out for tvlsmeeehs. It mkeas ssene besacue you lraen sothmnieg new about how the biran prsceeoss txet.

Ineedd!

gibberish is still not english. nor is it readable.

It is Engsilh. Only the order of the letrets is caehgnd a bit. It clairnety isn't a dffeneirt lagugnae. Aslo, you sluohd be albe to raed it if you raed the txet fsat. It's atalcluy more diiclffut to ustennrdad if you read it solwly wrod for word.

Excellent - very interesting.

And it's true - the faster I read it the more words I understand. It feels something like clearing a jump on a bike - the faster you go, the more likely you are to make it.

I'm sure that the brain relies on context in a big way to discern speed read words.

Blacknad.

this is a very interesting way of winning an arguement. make your post in such a way as to make it unreadable by the majority of poster. how can they argue with you, without a translator.

"interesting how the non-dyslexic mind automatically sorts it out making it legible"

I am dyslexic and have no trouble reading it (well no more than usually) then again maybe I am just used to letters looking a bit jumbled to me from time to time LOL

This is amusing.

It reminds me vaguely of an article that was floating around a few years ago about some research done at Harvard. (I think it was exaggerated, but there were some true elements, IIRC.)

Very good but a little long, Count.

I had friends of mine send emails in that fashion and I was surprized to see that I could read them in spite of the jumble.
The key is in the first and last letters.
jjw

All right, it wins a reprieve. But in future lets stick to Standard Modern English, please.

Amaranth