The end time field is a handy one. By default the curve will stop once you reach the final control point, but if you set the end time farther out it'll extend the curve to the specified time. You can also set it earlier to get a shorter roasting plan without changing the shape of the part you're keeping. Useful for the case of deciding you want to try roasting something a little lighter or darker.
A little preview of a thing I've been working on lately. The layout and visual design is far from finalized (there's no styling on the HTML at all) and there are at least a few more features that I want to add before I upload this somewhere, but the core functionality works. You give it some control points, it produces a plan that you can roast a batch of coffee against. Below the bottom of this image there are a couple buttons for exporting the generated plan either to CSV or Typica's XML.
@gnomon The cat seems to be fine. Currently sitting next to me.
Put in the export to Typica feature on the web thingy I've been working on. This can now officially replace my old C++ utility, though I still intend to add at least a few more features before I put it online for others to use. The next big one would be SVG generation so people can eyeball that the plan is what they had in mind before exporting that for use.
The latter seems like the easiest option and would be handy to have at events where there may or may not be a computer hooked up to the roaster at all and if there is it's probably not set up in a way that's convenient for taking your data with you.
The other option, which I haven't seen before, would be to have a timer update the time field continuously and raise the temperature value with each insertion so you could just hit the button when the temperature gets to the next value. No typing at all.
I'm kind of tempted to add a couple of live logging features, but if I start adding stuff like that I'll definitely need to add some styling to avoid confusing people with more advanced features. One option would be an auto-increment on the time field (so once you put in a point it'll advance that to the next 30s and focus the temperature field so there's just one number to put in when it's time).
I could throw together an API endpoint to generate a PDF and serve that, but I very much don't want to.
Does anybody happen to know how good/bad support for CSS for printing is these days? For the on screen thing I'm fine with unstyled defaults (though I'm not opposed to something that looks a little nicer), but I'd like it if people could print the page and get something that's arranged to fit nicely on a page with space in the right places to write notes on and without the stuff that doesn't make sense to include in a printout.
Last night I tackled the heaviest math of my current project. I already had a working C++ implementation so it was just a matter of porting it over to Javascript. Also did a tiny amount of quality of life features. It still looks very 90s (I may keep it that way), but slightly more than the absolute minimum needed to be useful is done.
Author of Typica software for coffee roasters.