Re: terminal only for two weeks

D <nospam@example.net> writes:

Brilliant! You are a poet Mike!

I'm doubtful that poetry can be done in Perl.  Maybe free verse in
Lisp.

Frogfind.com was a great start! I would love to have some kind of crowd
sourced html5->html1 - javascript - garbage script.

Do note that Frogfind delivers URLs that send your click back to
Frogfind to be proxied.  I assume that's how you get de-enshitified
pages in response to clicking a link returned from a search.

Here's a curiosity:

Google also sends all of your clicks on search results back through
Google.  I assume y'all knew that.

If you search for (say):

      leon "the professional"

you get:

https://www.google.com/url?q=https://en.wikipedia.org/wiki/L%25C3%25A9on:_The_Professional&sa=U&ved=2ahUKEwi [snip tracking hentracks/data]

Note that the "real" URL which Google proposes to proxy for you
contains non-ASCII characters:

   en.wikipedia.org/wiki/L%25C3%25A9on:_The_Professional

Wikipedia does *not* *have* a page connected to that URL!  But if you
click the link and send it back through Google, you reach the right
Wikipedia page that *does* exist:

   en.wikipedia.org/wiki/Leon:_The_Professional

AFAICT, when spidering the net, Google finds the page that *does*
exist, modifies it according to (opaque, unknown) rules of orthography
and delivers that to you.  When you send that link back through
Google, Google silently reverts the imposed orthographic "correction"
so that the link goes to an existing page.

Isn't the weird?

Try it.  Copy the "real" URL from such a Google response, eliding
everything before (and including) "?q=" and after (and including) the
first "&", paste it into your browser.  Wikipedia will politely tell
you that no such page is available and offer search suggestions.
Revert the non-ASCII "e with a diacritical mark" to 'e' (mutatis
mutandem for similar Google "hits") and it will work.

I also wondered if another approach might just be to take the top 500
sites and base it on that? Or even looking through my own history, take
the top 100.

Now there's a project suitable for AI: train the NN to treat a response
containing stuff you don't want ever to see as a failure.  Grovel
repetitively through terabytes of HTML and finally come up with a
generalized filter solution.

Due to the bad development of the net, it seems like a greater and
greater part of our browsing takes place on ever fewer numbers of
sites.

--
Mike Spencer                  Nova Scotia, Canada