Thanks for linking! There seem to be similarities, but looking (very) briefly at the DolphinScheduler these are the potential differences:
- number of contributors :D
- titanoboa can process even a potentially cyclic graph
- in titanoboa you can write step functions directly in high level programming languages such as clojure and java (so not just bash or python) and deploy them directly during runtime
- the clustering setup in titanoboa is master-less
- titanoboa does not have such direct integration with Spark as it employs some map-reduce patterns internally
But all-in-all I have to say that DolphinScheduler seems quite nice! Also would have to compcomplement it on the nice documentation (again, just briefly skimming through it).
It's just rare to get them on HN, because it's a nightmare to go through their docs and they're usually not even attempting to write their code in English. Basically unusable for all intents and purposes, even if it were quality software.
> Basically unusable for all intents and purposes, even if it were quality software.
Only if you don’t have anyone who can’t read Chinese on your team. Also most repositories are not documented at well or at all anyway so language hardly matters.
yes, most repositories arent documented well either, thats definitely true and was part of my point, really.
how are you going to figure out why you're encountering a bug if not even the code itself is written in english?
its fine for learning repositories or simple toy projects, but if you actually want your code to be used... please use the world language. (and no, english isnt my native language either)
A decompiler can't pull contextually appropriate variable and function/method names out of nowhere (not to mention comments), which is the big roadblock when reading foreign-language code.
That is, you're just as likely to be able to follow foreign-language code as you are decompiled code. Either way, you've basically thrown out all the documentation and swapped out all the names for gibberish.
I was just reading up on Broadway, written in Elixir, (https://hexdocs.pm/broadway/Broadway.html) that provides the fundamentals of batching/job control. It’s by the creator of Elixir and is based on 7 years of libraries in the area so the fundamentals are pretty well honed.
> Dagster is a system for building modern data applications.
> Combining an elegant programming model and beautiful tools, Dagster allows infrastructure engineers, data engineers, and data scientists to seamlessly collaborate to process and produce the trusted, reliable data needed in today's world.
Two paragraphs, communicating zero bits of information. I wish Github repositories, of all places, didn't contain such noninformative copy.
easy to use.