Login
You're viewing the dmv.community public feed.

Replies

  • 💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Face Thumbchrisp@cyberplace.social
    Oct 16, 2026, 12:16 PM

    @cstross Amazon and Apple could easily detect this. Presumably the Unicode matching works the other way too, search for the bad title and you would find the good title, then see if the author/publisher matches the existing, legitimate one, if it doesn't then flag it. But Amazon/Apple might miss out on a few hundred dollars, so we can't have that.

    💬 0🔄 0⭐ 0
  • robrykrobryk@social.wuatek.is
    Oct 16, 2026, 12:40 PM
    @cstross What are they avoiding by using a homograph? From your description it seems that author name matching for search and for whatever fraud detection mechanism works differently. Do you know why?
    💬 0🔄 0⭐ 0
  • Charlie Strosscstross@wandering.shop
    Oct 16, 2026, 1:07 PM

    @robryk What they’re doing is impersonating the author’s identity on the storefront where the customers see it, but creating a separate billing account that the pirate uses for receiving payments. The book title is the same, but because their book is cheaper many customers buy it instead of the real one.

    This process is semi-automated—someone did this to "A Conventional Boy" in its launch week.

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Charlie Strosscstross@wandering.shop
    Oct 16, 2026, 2:24 PM

    @robryk So they can set up a merchant account under a name the author/their publisher won't notice and stealthily divert funds to it.

    💬 1🔄 0⭐ 0
  • robrykrobryk@social.wuatek.is
    Oct 16, 2026, 2:31 PM
    @cstross Ah, and people search by titles, or the search is better at matching homographs than what the publisher's regularly doing.

    To be fair, this surprises me a bit: if the publisher has to do something active to find impostors that use the exact same name, I'd expect expanding that to names that look the same to be worthwhile and feasible.
    💬 0🔄 0⭐ 0
  • Charlie Strosscstross@wandering.shop
    Oct 16, 2026, 2:34 PM

    @robryk Search-by-title generally looks for synonyms and close misspellings (the retailer wants to maximize sales). But you *don’t* want to do approximate/fuzzy matching on your supplier names in your inventory and accounts payable database!

    I emphasize: this problem isn’t the retailer’s direct fault, any more than a SQL injection attack is. They should fix it, but the problem exists because of criminals.

    💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Charlie Strosscstross@wandering.shop
    Oct 16, 2026, 1:08 PM

    @BashStKid I’m not sure it’s still happening—I got hit by it 3 years ago, it’s possible they tightened things up since then.

    💬 0🔄 0⭐ 0
  • davecb 🇨🇦 :heart_transgender:davecb@hachyderm.io
    Oct 16, 2026, 1:28 PM

    @cstross

    In principle you can find those trick-homograph names.

    The Unicode folks have a TR on Unicode Security Mechanisms, which defines skeleton functions that find confusable characters in identifiers: unicode.org/reports/tr39/

    There are some python tools, like sr.ht/~valhalla/confusable_hom A google search found quite a number of these.

    A scraper for amazon is at github.com/omkarcloud/amazon-s, which in principle can get you the information to process, and which could host a checking mechanism.

    Not easy, but reachable, possibly by an association or group of authors. (:-))

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0