pfalcon

Very nice, thanks Jeff! Let me first elaborate why I'm asking. I started a project some time ago which in spirit is similar to i2cdevlib, though I decided to shoot for generality instead of consistency. Well, basic idea was to separate register metadata from another project (I hope to post about that too), to make that data more reusable. The format of metadata is not stipulated, in the basic form it would be just as C header with defines. Of course, I'd aim a bit higher, for some more structured metadata format, and well, did my homework to find some existing open-source solution and base on that instead of rolling out my own format. Well, I did find something to base on, and went with that (already extended format a bit, etc.). Format is based on YAML, which I'm pretty happy about, as it's more human readable/writable. You can see sample here: https://github.com/pfalcon/libperipha/blob/master/arm/cortex-m/lm4/gpio.yaml . The usecase both the original author worked on and I was interested in first of all is wrapping Cortex-M MCU register definitions. Note that I don't shoot for complete coverage - the idea is to describe what's currently needed, and then add more stuff on demand (for example, as contribution from other folks). So, I hope that explains my interest to the availability and format of your metadata. I don't want to start with proposing to work on converge formats or something (especially that I have more priority proposals ;-) ), but IMHO we should at least be aware of each other's efforts, and preferably make sure that formats are general enough to convert between them. libperipha's format should be general enough already and open to extension, and as far as I can tell your format doesn't have I2C-specific parts either. All in all, I'd appreciate your comments / some background info why you chose specifically this scheme (like, is it based on some existing format), what do you think of YAML as metadata syntax, etc. Thanks.

http://www.i2cdevlib.com/ front page mentions: Are ther plans or intentions to release this metadata as open-source, to allow community to develop even more innovative usages for it? (As a rough example, it would allow to provide bindings for more languages, starting with raw C.)