Besides the short names, common blocks and computed gotos were the bane of my existence maintaining primordial Fortran. My first and intern project was translating a punch card era (early 60s was the best bet) code base that had been textified and could compile in f77. Fortunately I didn’t have to really deal with any strings
You should check out the negpy project. It in my view has the best inversion process. It claims to emulate the whole analog print chain in the inversion process. They have a fold works cited.
It also automates most of the process. It works with SANE, some other legacy scanners, and camera scanning. It can even do trichrome scans automatically.
I've played with this lately. I'd pair it with something like VueScan as the scanner setup is a bit lacking still, but overall it is really impressive. The only difficulty I found is coming into the terminology it uses fresh - it's quite different to most other photo software in that regard.
How lately? They have been releasing quite frequently, and have worked a bit to clean up the UI and workflows after adding a ton of new features. I've been using it and it's working quite well for me after version 0.60-ish
About a month ago I think? It didn't have enough options to run the Coolscan 4000 ED I have lying about, at least not reliably, and I don't think it had any batch options (might be wrong about that). I don't really mind going through VueScan first but I'll take a look again soon and see if that area has been improved.
CoolScan support has improved a lot in the last month or so. I don't have one so I cant comment in practice, but its a very active project and I'd definitely encourage you to submit bugs if you have a coolscan.
Ooh great I'll check it out. I have a couple of them actually, although I couldn't seem to get one working through the complex series of adapters I had, so I'm gonna set up a new machine with an actual FireWire card and reassess it then. I do at least have one that works in VueScan so I can at least provide some feedback (it's a modded 4000 ED that does whole rolls).
Funnily sub-zero fridges are considered the samsung of the semi-commercial brands. They prioritize things like customizable doors and easy to break but consumer loved features like ice makers and water dispensers. You will never see sub-zero in actual commercial kitchens.
The value of writing isn’t always to communicate new information. It’s often to align everyone’s assumptions. For instance, when I say casually to a colleague or an agent “this change will require a db migration” they understand it’s to my teams primary application database. If I submit a design doc to a company wide review which database is changing is critical information. If the 300 bits were truly enough your agent or junior engineer would implement the wrong thing correctly as they often do.
Clarifying which database you're referring to is exactly the sort of thing I'm talking about when I say "new semantic information". It should be communicated, and you can't trust an LLM to choose the correct database, you need to specify that yourself.
What I'm talking about is if you add "database foo" to your prompt, the LLM may then add text describing what that database is, where it is, etc. But that's not new information, it (hopefully) already exists in your team's public docs, slack convos, etc. You should just say "database foo" directly to your reader, and if they want to learn more about that database, they can do that themselves, or you can give pointers to them based on what you consider important.
> say "new semantic information". It should be communicated, and you can't trust an LLM to choose the correct database, you need to specify that yourself.
But I can most of the time, and correct it if it chooses the wrong thing. That's the whole reason LLMs are faster. Its the reason we can give a paragraph prompt and get a kLOC PR back but only need to correct about 5% of it.
> The actual risk that the author does not even broach upon for investors... the thing that will actually torpedo this massive investment are the open source open weight chinese models that commoditize the entire endeavor.
OpenAI and Anthropic investors yes, however open weight models are good for cloud providers. They can turn the two large customers into direct ai services that can be spread across many customers and reduce the cloud providers overhead on ai services.
