7 ms·
I now noticed that you can read the page if you hit escape fast enough: https://www.nytimes.com/programs/better-reader/day-1 https://www.nytimes.com/programs/b
by FreeHugs 7y ago
I now noticed that you can read the page if you hit escape fast enough:
https://www.nytimes.com/programs/better-reader/day-1 https://www.nytimes.com/programs/better-reader/day-1
When you visit it, it shows the page for a split seconds. Then redirects. By hitting escape in that split second, the page stays on screen and the redirect does not take place.
Modern developers :-) Implement everything clientside in Javascript.
It's kind of a fun game. Especially since the page then has no ads at all. All newspaper pages should look like this.
- weinzierl 7y agoSpiegel Online used (or uses, idk) a blurring overlay to hide their payed content. My first thought was "That'll be easy!" and I removed the overlay only to find that the underlying text was obfuscated in a way that made it look like the original text through the blur. Well played Spiegel! Still, I suspect the substitution could at least be partially reversed but I never tried. What makes it more complicated than a usual substitution cipher is that it can be arbitrarily lossy because it isn't meant to be deciphered. As an extreme case they could just replace all uppercase letters with X and all lowercase letter with x - for example - and no information could be gained while still looking similar when blurred. Luckily it's not what they did, the substitution looked more complicated. Another reason to make me believe that it's reversible is that I think I remember David Kriesel (the "Lies, damned lies and scans" guy) once hinted that he did it. Anyway, deciphering it would be a nice Sunday afternoon entertainment, I guess...
- pintxo 7y agoThey did in fact use a Caesar cipher [1], a quick calculation on letter frequencies gave a strong hint. The rest was looking up the shift value and coding a js bookmarklet [2] to "decrypt" the text. It had some issues with certain German Umlauts, but was not enough of an issue to fix it. Looks like they now longer include the full text publicly. So they seem to have improved. [1] (German) https://andreas-zeller.blogspot.com/2016/06/spiegel-online-nutzt-unsichere-casar.html https://andreas-zeller.blogspot.com/2016/06/spiegel-online-n... [2] javascript: document.querySelectorAll('div.obfuscated-content')[0].parentNode.classList = [];var cc = (s, c) => s.split('').map(s => /[^\s]/.test(s) ? String.fromCharCode(s.charCodeAt(0)+c) : s).join('');var dn = (n) => { if (n.hasChildNodes()) {Array.from(n.childNodes).map(dn);} else if (n.parentNode.nodeName !== 'A') {n.textContent=cc(n.textContent, -1) }};document.querySelectorAll('p.obfuscated').forEach(dn);document.querySelector('.lp_mwi_payment-method-wrapper').parentNode.parentNode.remove();
- pintxo 7y agoOthers [1] still seem to use this technique. [1] https://www.faz.net/aktuell/feuilleton/debatten/klagenfurter-kunststreit-was-soll-ein-wald-im-stadion-16377833.html https://www.faz.net/aktuell/feuilleton/debatten/klagenfurter...
- ThrustVectoring 7y agoI usually just download the URL in the terminal and open it locally when pages do that nonsense.