Programming Models for Distributed Computation
1–10 of 30 posts
Re: Programming Models for Distributed Computation
#2The content seems shallow, the topic is extremely broad. Overall I do not feel this topic can be described really well.
Re: Programming Models for Distributed Computation
#3Linear Logic / Session Types seem interesting. I've poked around some of the papers, but I'm not sure how to approach the subject yet.
https://homepages.inf.ed.ac.uk/wadler/topics/linear-logic.ht...
Re: Programming Models for Distributed Computation
#4The content seems shallow, the topic is extremely broad. Overall I do not feel this topic can be described really well.
I think "shallow" is being a bit uncharitable. The book here doesn't dive into ditributed systems theory very deeply but it gives a very rich overview of real industry tools used over the past few decades with plenty of context around why technologies are being used or discarded. They go into code examples of things like CRDTs so its not like its just prose.
One thing that feels off to me as a backend engineer that deals with distributed systems in practice is no treatment (AFAICT) of containers/virtual machines and how tools like kubernetes are providing the primitives that the distributed languages of the past provided. In general I think chapter 4 doesn't have a cohesive narrative on its treatment of languages for distributed computation. I know this is messy topic but I think more could be done to tie things to current art.
If I were to make a prediction, there's general trend towards protocols (mostly network level above layer 4) being the actual thing of interest for distributed computation, not single language runtimes. Wire-compatible DBs like AWS aurora and MySQL are an example, but also I think things like container orchestator's liveness and readiness checks over HTTP. As more primitives rally around the same core concepts, I think the general trend will be to go from conceptual similarity, to client library compatability (e.g. identical interfaces a la OpenTracing Lightstep, Jaeger) and eventually wire/protocol compatibility. Language runtimes are too much of a moving target as there are other pressures at play (community, tooling, typing, yes even syntax) that go into language choices over a decade.
Re: Programming Models for Distributed Computation
#5The content seems shallow, the topic is extremely broad. Overall I do not feel this topic can be described really well.
I think "shallow" is being a bit uncharitable. The book here doesn't dive into ditributed systems theory very deeply but it gives a very rich overview of real industry tools used over the past few decades with plenty of context around why technologies are being used or discarded. They go into code examples of things like CRDTs so its not like its just prose. One thing that feels off to me as a backend engineer that d…
Yes, i was also a bit disappointed reading e.g. the reference e.g. about CSPs, it should at least link some other textbooks here. However, I admire the work being put into this. It would be great if all university teachers would make the effort to publish open access. A dream would be collaboratively maintained textbooks with strong single editors for readability and consistent style. Also there should be foundations like apache, etc that market such material and put up quality control, pay lectors where the publishers are largely failing atm.
Re: Programming Models for Distributed Computation
#6This is going to sound snobby to some people but I don't care. I really believe this.
The whole book is written in Markdown. This is a huge red flag for me and the book goes in to the "no" pile pretty much immediately.
Distributed computation is a complex and complicated discipline. Moreover if you're going to write a book and you don't even realize that Markdown is completely inappropriate it calls in to question your authoring skills.
At best the use of Markdown is lazy.
If you are an expert in distributed computation I assume that you're probably capable of using LaTeX, DocBook, *roff, or some other appropriate technical authoring system.
Re: Programming Models for Distributed Computation
#7This is going to sound snobby to some people but I don't care. I really believe this. The whole book is written in Markdown. This is a huge red flag for me and the book goes in to the "no" pile pretty much immediately. Distributed computation is a complex and complicated discipline. Moreover if you're going to write a book and you don't even realize that Markdown is completely inappropriate it calls in to question yo…
What does writing in Markdown have to do with the complexity of the topic? Heather Miller is an assistant professor at CMU and well respected and certainly an expert in this context. Clearly she believed Markdown was sufficient for her purpose and/or wanted to leverage the other benefits beyond just its simplicity. Did you see that book is being rendered on a site? http://dist-prog-book.com/chapter/2/futures.html
Re: Programming Models for Distributed Computation
#8This is going to sound snobby to some people but I don't care. I really believe this. The whole book is written in Markdown. This is a huge red flag for me and the book goes in to the "no" pile pretty much immediately. Distributed computation is a complex and complicated discipline. Moreover if you're going to write a book and you don't even realize that Markdown is completely inappropriate it calls in to question yo…
It doesn't come off as snobby or even pretentious, merely facile. The author has publications in distributed systems in reputed journals, is currently researching and teaching on distributed systems, and has worked with the Scala community for quite some time so at the very least she is competent and knows the subject.
Dismissing content because it's written in markdown without providing good reasons, such as quality of content or specifically how the presentation hinders proper communication is indeed quite lazy. If you want people to take you seriously, please try to be more precise in your criticism.
Re: Programming Models for Distributed Computation
#9lol hn never fails in it's shallow dismissal of an enormous amount of work: as of right now there's a comment on how this textbook is too shallow (I guess if it were a content marketing medium post it would be much better) and how it can't possibly be substantive because it's written in markdown.
man what a bastion of intelligent discussion this place is.
Re: Programming Models for Distributed Computation
#10This is going to sound snobby to some people but I don't care. I really believe this. The whole book is written in Markdown. This is a huge red flag for me and the book goes in to the "no" pile pretty much immediately. Distributed computation is a complex and complicated discipline. Moreover if you're going to write a book and you don't even realize that Markdown is completely inappropriate it calls in to question yo…
What does writing in Markdown have to do with the complexity of the topic? Heather Miller is an assistant professor at CMU and well respected and certainly an expert in this context. Clearly she believed Markdown was sufficient for her purpose and/or wanted to leverage the other benefits beyond just its simplicity. Did you see that book is being rendered on a site? http://dist-prog-book.com/chapter/2/futures.html
>What does writing in Markdown have to do with the complexity of the topic?
nothing. the person you're responding to just wants to feel very good about themselves without actually engaging with the material.