Ask HN: What are you working on? (February 2025)

That is super useful, especially for small stuff in home labs. I was thinking to go ProxMox + Gardener + Kubevirt route, but never have time to work on it.

I'm working on PayPerFax - https://payperfax.com/ - a simple online faxing service that lets individuals send faxes without committing to subscriptions or creating accounts.

The idea came from an interesting "tail wagging the dog" situation. I had an old domain (faxbeep.com) lying around unused which I'd been renewing for 20 years. When I finally built something there, I discovered users were more interested in a fax testing feature that I'd added as an afterthought - the abitlity to send a fax to our test numbers and see it appear on the website for 30 days.

That insight led me to dedicate faxbeep.com entirely to fax testing, and to buy payperfax.com for my original idea: straightforward, pay-as-you-go faxing. People still need to send faxes surprisingly often - for tax authorities, immigration paperwork, medical prescriptions, etc. The service charges a fixed fee for a 3-page fax (cover + 2 pages) with a bit more for additional pages, and you only pay if the fax sends successfully.

My current challenge is visibility - Google has essentially black-holed the site. Even exact match searches like [pay per fax] don't bring up the site. If anyone has SEO advice on how to climb out of this hole - I'd appreciate it!

Yes, that is a big part of the design choice for this language. I felt that circuit elements are usually defined in relation to other circuit elements and instead of explicitly stating connections, it might provide more understanding of the overall circuit if the user builds the connections by "walking" along the circuit graph. A good side effect of this is that you skip having to name every single component in the circuit.

I was inspired by skidl (https://devbisme.github.io/skidl/)'s python usage and did explore that direction. However, it felt strange and clunky to try and fit the circuit drawing concepts I had in mind into python via a domain specific language (DSL). The user would require a decent understanding of python and the DSL to create circuits. This would complicate my original aim of creating a language/tool that allows electronics engineers to focus on the circuits that they would like to design.

By creating a new language, I was not limited by python's syntax/grammar. I could craft the language closer to what was needed for schematic design and maintain simplicity. Overall, it was a great learning experience on language design and also using ANTLR for the implementation.

creating a AI generated data pipelines. building config data driven pipelines that enables domain driven data teams

zeitgeist observations regarding this:

haystack editor is a neat canvas based IDE

codesee recently got acquired by GitKraken (wish they'd sell individual licenses for function maps rather than only for enterprise)

Sometimes I wish I could code within an obsidan.MD canvas

I think a lot of developers generally want to code on a useful node graph IDE ---- the Gource time travel idea is interesting!

Would love to see your idea pulled off! I'm rooting for you