I remember using Pidgin in ~2009. A dozen chat networks, all on one app. Desktop software built with a native GUI toolkit. And, on top of all that: you could keep your chat logs forever. The world of yesterday.
There was a plugin called "Off The Record" (OTR) which would do a pk exhange and then send cipher text over the channel. It was rad. You could have e2ee over Facebook Messenger. When you opened the chat in the Facebook web ui, all you could see was the cipher-text.
Then Facebook started blocking 3rd party clients and Pidgin et-al slowly faded away.
I remember! I also used Pidgin OTR over the Facebook XMPP gateway. At some point Facebook started recognizing it, but not banning it: you could go to the web interface and you'd see "encrypted message" instead of noise.
Mark Zuckerberg hacked Crimson reporters (a Harvard newspaper) who were investigating him for Facemash. Mark Zuckerberg's company took people's API-facing emails that were in their profiles and replaced them with a facebook.com address. Mark Zuckerberg's company deleted years of his own correspondence. This in addition to things like Onavo
Yeah, I'm suggesting he go to prison for "doing whatever he wants"
I remember I had a plugin that let you change your profile picture each <x> time. And I seem to recall with ubuntu's notify-osd you could reply to your incoming messages from within the notification itself. I loved using Pidgin.
"Modern" mainstream IM is completely misserable. I hate having to use one-app-per-each-protocol for the sake of "security" and "features".
It's not rose-tinted glasses IMO. Aside from cross-device continuous chats (which weren't really relevant at the time) and maybe being harder to send pics (can't recall), Adium was a far better messaging experience than anything modern.
* You could theme it however you wanted to an obscene amount. I had it display all messages right after each other in a small font without any linebreaks and I've never been able to have anything like that since then.
* The dock icon showed the names of the last few people who sent you unread messages
* It integrated with the OS X phone book app so you could it would display a single "John Smith" regardless of how many chat apps (AIM, MSN, Yahoo, etc.) you had them on
* It was actually smooth and not clunky (unlike Pidgin at the time and maybe half of apps today).
I used Kopete with inline videos and a newspaper-like theme. It was amazing and beautiful. That under 256MB of RAM. Nowadays you would need 2GB to do the same.
And bear in mind KDE3 was considered the bloated DE, as XFCE (even the GTK2 build) could snappily run with 64 MB of RAM and maybe less with a light GTK engine (yes, choosing the GTK2 engine mattered a lot back in the day).
And, yes, choosing Pidgin and a light window manager such as Fluxbox/Openbox could make run machine run well with 64MB at really fast speeds.
Great nostalgic reminder! Multi-protocol clients like Adium and Pidgin offered unified messaging and features like persistent logs and customizable interfaces that modern apps often lack.
You can still use Beeper[0] and similar. The key issue with this type of application is that some networks have put more resources to detecting them and gotten more hostile to users of it - mostly those who tie ad revenue directly to messaging (although officially it's to avoid spam + detect compromised accounts).
I was surprised to see that Beeper actually has support for ‘local bridges’ that connect to services on-device (which reduces the risk of bans and removes Beeper as the middleman).
I was unsurprised to see that (at least with the local Instagram bridge), Beeper is extremely inconsistent with push notifications and sometimes has messages missing in the chat.
Sure, but unfortunately most people are now using iMessage, Whatsapp, Signal, Facebook messenger, and so on, and Pidgin can't connect to any of them AFAIK.
Interesting, there are apparently working plugins for both Signal and Whatsapp! Both appear to work as secondary devices. What would really be great is to have the possibility of using Pidgin as primary device...
From the same people, get Bitlbee with libpurple. IRC logging with your favourite client against everything supported by Bitlbee AND the Pidgin library.
You can connect from any OS with an IRC client. It's astounding and liberating.
I expect it's somewhere in the training data, but it's very unlikely to be salient. A few textfiles here and there in the ocean of the Internet is nothing. If Claude had memorized the walkthrough, it would have performed better.
Microsoft's revenue in 1990 was 1.18B when they launched Office, sold for one-time payments. Of course they're pushing people to subscribe now so they can get that sweet recurring revenue, but that business model sustained freaking Microsoft for about 30 years.
I'm not convinced by unsustainability arguments. Now, it could be that competing with FOSS makes it a lot harder to make money now. I'm sympathetic to that, inasmuch as I can be for someone who wants to sell what others are giving away. That would be challenging. But why is it suddenly impossible to sell software, when they was the common model until rental became popular a few years ago? What's inherently different now that let someone sell programs for decades but now it's just impossible?
buying - capital expenditure with amortization (and usually goes through a lot of approvals, centralized IT, etc.), subscription - expense, frequently decided upon and paid directly by the Line-Of-Business/dept. Expense is generally better, so it is chosen by business when possible (it is all very generic of course, and there are niche cases where situation is different)
That matches on the supplying side as subscription revenue is also generally better.
Fairly certain it'd be more sustainable than a $4/month subscription, which is a nuisance for anybody who'd actually want to pay for this.
$4 is targeting the hobbyist market. Within that segment, the tiny population of devs who'd actually be willing to pay for tools usually uses a large assortment of tools, and is not willing to pay a separate subscription fee for each.
5,000 seats x $200 price = $1,000,000 cash realized, right now, no debt financing needed for funding development, regardless if said dev shop leaves you.
5,000 seats x $4 = $20,000 month. That kinda pays the salary of a single FTE, and development ceases if customer runs away, only safe option would be to finance further development to insulate yourself from churn.
With that, he could invest it, finish the features, start work on 2.0 and charge and upgrade of $20 to all his customers when he has MSSQL, MySQL, pgsql, SQLite, Oracle. 2.0 could include cloud db’s.
Not everything has to be a subscription and if people continue to find value, they’ll upgrade if they think it worthy.
Subscriptions tend to be vampires on the wallet and are used to trap users into paying a developer indefinitely. I consider it hostile to users. The exception is if it’s a service that can’t run on my machine. Or a service that does back office for me.
There are also things like Dolphin actions/addons (I forget what they’re actually called) that you can add so you can do a conversion with a right click.
They basically consume the libheif command line tool so you install that as a prerequisite.
How did you get this to work? (See the config and did this before but never got ly to work on nixos, i migth have just been to early i guess)
Ive been using LY on arch for a while, and it is great looking, when it works, but it has lots of issues.
Mostly related to multi monitor, kernel warnings filling up the screen, and for all being just sligthly broken. But then again probably user error.
In fairness to Emacs, this is a bit sour grapes on my part!
I have tried to go fully into the "Emacs mindset" (org-mode for everything, multiple pages of custom hydra keybinds etc.) a number of times and I always bounce off. I always feel there is some activation threshold that if I could cross it, I could enter editor nirvana.
I used to joke that the way I use Emacs is I open it, give the empty buffer a very meaningful look, C-x C-c, and open VS Code.
For whatever it's worth, I think in 2025 with good LLMs, Emacs is actually bliss. Even as a true believer, I would regularly think of customisations, and then sigh at the effort and not bother. Now, I just get an AI to help me write the Emacs Lisp which not only teaches me new things, but also gives me (in seconds) an upgrade to my productivity which will last forever. Not only that, but I am using LLMs in my editor to help write code to make using LLMs in my editor even easier, so I feel like I've simultaneously crossed two thresholds.
My story is a lot like yours, except swap the two editors. I decide I'm really gonna try Visual Studio Code this time. Everybody uses it, it's become the default editor for like every recent programming language... it must be better than what I'm using, right? Fifty million Elvis fans can't be wrong!
And then I fire it up and... it's not compatible with my muscle memory. Plus I can't just pop open a buffer and morph my editor into what I need for the task in a language I like. (There is considerable rigamarole involved in writing a Visual Studio Code extension; I tried.) I can't work with buffers the way I'm used to, it doesn't indent the way I'm used to... and unless I'm willing to limit myself with VSCodium, it's spying on me in a way I consider hostile. So I put it away and get what I need done in Emacs. I must've been through this cycle like, six times.
reply