BTW. This, actually, is the first comment I got; this blog is brand new. Thank you.
260 karma · joined November 22, 2012
BTW. This, actually, is the first comment I got; this blog is brand new. Thank you.
I've braced myself for ANY feedback you can send my way, so go ahead. :) I don't work to work on this in isolation.
I'm personally quite comfortable with the API so far, the only real doubts is error handling, i.e. IncTemperature from the PDF can optionally return an error and if that happens, it kills the session (state for an open browser tab) and restarts it with no state. Because errors need to be handled as regular state in LiveView. But it may be counterintuitive for Go developers. If anyone is interested, I can go into more details on this one.
But I agree AFTER it got moved to Tooploox Github it looks kinda self-serving.
You can: 1. Create bash-based tasks similar to what Makefiles do.
2. You can parametrize these tasks AND templates using YAML files such as values.oya (they can be encrypted).
3. You can share/reuse your scripts by just pushing them to Github and tagging using a version and then import them into your Oyafile (`oya import github.com/bilus/mypack`)
Just look at Go channels or at reactive programming patterns. If applied to the right problem (i.e. a problem they help solve), they allow you to solve the very problem in a very concise and expressive way.
Having such patterns as a part of the language just makes them more popular and reusable.
And as far as your point, don't you think it's easier to spot a 5-line pattern than a pattern that is spread out over several classes and 200 or so lines of code? It all boils down to this: being able to see patterns or abstractions if you know that and having a common language across the team. Functional 'patterns' are just (arguably) more succinct than object-oriented ones (and it comes from someone who has been programming in C++ and Ruby for nearly 20 years).
Me too BTW. ;)
Experiments with revenue are very important too though.
What do you think? Should we do revenue right away or retention/viral aspects first? I sometimes have doubts about this.