Continuing with the 'the program is the database' theme from last month, I made a clojurescript dialect (called preimp) where the only source of mutable state is the source code itself. It works well enough for spreadsheet-sized datasets:
I have much more interesting interactions planned. Hopefully for next month.
Doing this in clojurescript makes it easy to get things off the ground but creates a very low ceiling for how high-quality an experience I can make. So I'm thinking of this as a two-pronged assault: on the one prong I'm trying to make a language and programming environment that is ideal for data soup problems, and on the other I'm hacking together something awful in clojurescript to see where I have to go.
I wrote some musings on how data is represented in programming systems:
HYTRADBOI is clocking in so far at 169 attendees and 35 speakers.
If this slow trickle of sales continues we're looking at maybe 300 people total. More if there is a last-minute uptick. Perhaps a little more success than I was prepared for :S
I've turned off the office hours link - now that the weather is drying out I'll be either coding, writing or climbing pretty much 100% of the time.
I'm still open to unsolicited email though - async communication is much easier to fit into the gaps.
I'll also be in NYC May 5-13 and May 20-21.
- Building data-centric apps with a reactive relational database. Always exciting to see smart people working in this direction.
- Cue is one of the many 'programmable json' languages popping up. Notable for noticing that the override functionality replaces the need for functions. I think this might be an interesting set of semantics for an end-user programming language - rather than requiring users to structure abstractions up front, just build copy-paste-override into the language semantics.
- Technical Dimensions of Programming Systems. Trying to enumerate axes on which programming experiences can be compared. You have to have a map of what is known before you can go exploring.