Hi Jean-Sebastian, Honestly I have not had a chance to play with Seaport yet (Hope to get to it soon). Someone else also asked about maintenance here: https://github.com/substack/seaport/issues/72
Looks like he's still waiting on a response. I imagine you could take a REST HATEOAS approach to WRT communicating what services are related to / referenced by a particular service: https://spring.io/understanding/HATEOAS Superagent is a pretty impressive node REST client, in case you wanted to experiment. I'm also working on a Maven like build system for Node called VerdiJS. https://github.com/verdijs It will use self contained modularized Gulp based tasks for compiling ES6, Coffescript, etc. like this one: https://github.com/verdijs/verdijs-task-clean And utilize the same type of standard directory structure that Maven uses. https://github.com/verdijs/verdijs-pli Cheers, - Ole On 10/21/2015 01:00 AM, Jean-Sebastien Delfino wrote:
Hi Ole, Seaport looks nice and could maybe become part of a Node-friendly solution. It looks like a registry like many other registries out there allowing me to lookup a service name@version and get its host:port, but pretty light weight. Do you know if Seaport would also allow me to store the fact that a service X references service Y? (as that's really what I'm looking for with SCA references bound to services, locating service Y is only a small part of the story...) Also is Seaport still actively maintained? Thanks! - Jean-Sebastien On Sun, Oct 18, 2015 at 11:07 PM, Ole Ersoy <ole.er...@gmail.com <mailto:ole.er...@gmail.com>> wrote: Hello Jean-Sebastien, You may also want to have a look at [Top 10 Browserling Inventions](http://www.catonmat.net/blog/top-10-browserling-inventions/). I think you would be interested in the Seaport Service Registry, Ploy, Airport, Upnode, and Bouncy. Cheers, - Ole On 10/19/2015 12:25 AM, Jean-Sebastien Delfino wrote: Hi all, It has been a while... Today I was reflecting on what I've been doing in the last two years, mostly micro-services on Node.js, and I'm starting to think that the original ideas behind SCA and Tuscany may be useful to me again. So you may hear a bit more from me on this list again in the next few weeks... My new world is very different from the world we initially created Tuscany for: Node.js, Javascript everywhere, isomorphic Web apps, simple REST 'services', simple middleware and databases, and not much technical complexity getting in the way of writing business logic. Many of the issues we were trying to address with SCA like multi-language, multi-protocol, complexity of the JEE platform and WS stack, weird objects requiring injection etc, don't exist anymore in my new world. That's great as developing Web micro-services has become really easy! So easy that I have so many micro-services in my apps now that sometimes it gets a bit hard to keep track which service calls which, what's that service address, what I need to change when that service moves or gets updated, or what's involved when something goes wrong and I need to find which service broke. That's a serious problem, and something that made me think about SCA and Tuscany again. Despite all the greatness of Node.js and REST and micro-services, I'm probably still missing some kind of assembly model like we had with SCA. Something that would model my app as as an assembly of micro-services. Something that would allow my services to reference each other without having to update environment variables all over the place with their addresses. Something that would allow me to understand that a service broke because another service that it references is currently down. Something that would provide a description of my service call graphs for debugging for example. Right now, it's really easy for me to develop micro-services and wire them together, but I don't have a good way to model that wiring. Maybe what I'm looking for is a small subset of the original SCA concepts: a description of my app as an assembly of services, Javascript friendly, simple and lightweight, declarative but programmable, and distributed and dynamic as my services need to move around to scale out or when a Cloud region goes down. So, I'm going to spend some of my spare time on this, evenings and weekends, and try to put together a new variation of Tuscany for Node.js. I'd like to figure out if that good old SCA can help me again with my little micro-services issues. I'm thinking about calling that new variation of Tuscany 'Tuscany.js', and maybe put it in a new 'js' sub-folder in the Tuscany repo besides the existing java and cpp folders. I'd love to work on it with other folks in the community if they're interested! Thoughts? - Jean-Sebastien
