i wrote a program that counts requests by IP and frequency.  That and some
fancy code to teduce server load and the bots are blocked fairly well.  I
capture the attempts and lot them so I can see where and how theyre doing
it.  Inhave noticed many bots use the same exact probe progression, there
must be a popular scraper bot you can download
b


On Tue, Jul 7, 2026 at 6:58 PM Dennis Boone via cctalk <
[email protected]> wrote:

>  > I don’t care why, I’m not changing my browser configuration to
>  > accomodate whatever it is they think they need to make it work, but
>  > it does mean that more and more of the web has become inaccessible
>  > for me.
>
> As a guy who has to keep a variety of web sites running, I can explain
> the why:
>
> 1. Spiders and robots, many driven by the current AI craze, have
> increased in number, mendicacity, lawlessness, and aggressiveness by
> orders of magnitude.  If you're running a web site that provides content
> via some sort of dynamic mechanism (content management system) that runs
> code and does database queries to generate each response page, the new
> level of offered load squashes your server into a greasy spot.
>
> 2. These spiders (a) lie about who they are, pretending to be real
> humans; (b) don't obey robots.txt; (c) retrieve the same page multiple
> times at the same instant, sometimes using different user agent strings,
> to try to be sure they get the page in the face of increasing server
> hostility to their bullshit.
>
> 3. These bots don't care about request rate; they just expect you to
> deal with it.  Will the VPS where you host your ccmp Wordpress install
> keep up with dozens of page requests per second?  Because if the slop
> machine gets interested in something, those requests will come, and will
> probably flatten you.
>
> 4. Search and (especially) AI companies have had to resort to making
> end user machines fetch the content they want.  Smart TVs, phones,
> tablets, and other devices are increasingly being used.  These devices
> live in IP address space that's assigned to DSL, cable modems, cell
> companies, etc.
>
> ... and the difficulty of surviving the onslaught:
>
> 1. Because these requests come from where real users live, blocking by
> IP address blocks the real users.
>
> 2. Because the user agent strings are fabricated to be indistinguishable
> from real user browsers, blocking by them blocks the real users.
>
> 3. Blocking by behavior has some promise, but you really need insight
> into behavior across a wide swath of the internet, unless you're that
> big yourself.  I.e. Cloudflare and Google and Amazon have a chance.
> Hobbyists don't.
>
> For a dynamic web site, that basically leaves putting some sort of
> challenge in front of your site that humans can solve easily, but the
> bots can't (or don't want to bother), coupled with broader intelligence
> on trends across the 'net.  That looks, unsurprisingly, like Cloudflare
> and similar services.
>
> Unfortunately, those kinds of challenges, and the interface between them
> and the protected website, is largely dependent on javascript and
> cookies.  Grumble.
>
> Welcome to late stage capitalism.
>
> The future, for those of us who hate the technologies, budgets, and ick
> factor required to survive the post-web sewer, probably lies in static
> web sites.  Think tools like Hugo, Jekyll, Pelican.  Edit the page on
> your local machine.  Run the generator.  Synchronize the updated flat
> HTML pages to the server.
>
> Web servers deliver flat files with remarkable efficiency.  sendfile(2)
> is damn fast.
>
> Additional upside: it's hard to hack a web app that has no code running
> on the server.
>
> De
>

Reply via email to