I'll take "problems we solved correctly in the 80s" for $100, Alex


It's 2020 and developers still think "reverse chronological order, scrolling until I see something I vaguely remember having seen before for half a second" is the way to go. Good job everybody.


Proper threading, a real editor, post scoring, and kill files: trn was the pinnacle of user interfaces for keeping up with the social media feed.

Previously, previously, previously, previously, previously.

Tags: , , , ,

27 Responses:

  1. Volker Stolz says:

    But can’t shove targeted ads down your eyeballs, so, no you can’t have it.

  2. mattl says:

    Started playing around with trn again today, strangely.

    • Eli the Bearded says:

      github.com/acli/trn will get you a version that really understands charsets.

  3. margaret says:

    that's some fancy network cabling.

    • jwb says:

      The vt100 running flat out could do 19,200 bps. You can probably achieve this line rate over lamp cords.

      • db48x says:

        Yes, in fact you can cheaply achieve that rate over a lamp cord while the cord is still powering a lamp.

  4. Ahh, how I long for the days when software was developed for the user.

    • MattyJ says:

      And usually by the user.

      • There was something about the 8 bit days and writing your own stuff, but I was happy enough when other people would sell me software and actually care that it worked. And, you know, when computers were simple enough that it was possible to produce software with only a couple of bugs.

        • Rob Green says:

          Yes, clearly the reason that social media is totally unusable and that one can't even choose threaded display, inline search, or quickly pan to a specified date period is because programmers have to spend all their time coding around bugs, not because they are tasked with making the sites a time suck.

          • Oy. The sarcasm, it burns. I was more thinking that I can't get a decent word processor anymore, but you aren't wrong about the root of a different issue.

  5. Kyzer says:

    Let's add to that misery. Many people now enjoy communicating through memes, aka 100KB+ pictures of text. Why do they do it? Because it affords them some way to make their text stand out and you give them more attention. How do you find it again? You don't, just go back to the fresh stream of new memes and give them more attention.

    I imagine that someone's research project right now is OCR transcription of memes.

    • Jim says:

      Regular expressions for memes.

    • tfb says:

      This reminds me of one of Erik Naggum's signatures:

      a good picture may well be worth a thousand words, but on the WWW, even bad imagemaps cost tens of thousands of words.

      Except, now we will also expend hundreds of billions of machine cycles turning the pictures back into words again, unreliably.

      But it will be 'AI' so it will be progress.

    • someguy says:

      I imagine that someone's research project right now is OCR transcription of memes.

      That's... not a research project. That's (sadly) a real thing. Since memes are usually a common font like Impact and the text is high contrast (to stand out on the image), an OCR engine like Tesseract can pick it up with a decent degree of accuracy. One implementation was a popular bot on Reddit for a while.

      • jwz says:

        OCR has gotten good enough that Facebook (and also Twitter?) do it automatically on every uploaded image, for whatever nefarious purpose.

        Among other things, they prohibit you from running ads that have too large a ratio of text in them.

        We also constantly run into this problem with the restaurant ordering sites who have a filter enforcing their policy of "no text in pictures of food", which makes it impossible for us to show product images for, say a can of Red Bull, or DNA Lounge brand Whiskey".

      • The research project, naturally, is generation of memes:

        We introduce a novel meme generation system, which given any image can produce a humorous and relevant caption... The system uses a pretrained Inception-v3 network to return an image embedding which is passed to an attention-based deep-layer LSTM model producing the caption...

        Testers were able to figure out that the memes were computer-generated only about 70 percent of the time.

  6. tobias says:

    bravo. fab. your site works from my fullscreen shell without anti aliasing on lynx. maybe forgoing video and pictures is how escape UX wheel reinvention. this trn looks great.

  7. Michael says:

    I always used nn and thought it was the perfect way to efficiently browse large volumes of newsgroups. nn stood for "no news is good news".

  8. Jeff Bell says:

    The $100 category is also defunct.

    Since 2001 the minimum is $200.

  9. metarza says:

    Can I just take this moment to say that terminal is a thing of beauty. And the case is so clean and bright!

  10. Jyrgen N says:

    I am a member of the VT220 generation myself. While I used rn first, only days later the TA of my first Unix course pushed^Wpointed my in the GNUS direction -- and with that, also in the Emacs direction. Still use Emacs, still use Gnus, after 31 years.

  • Previously