Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

sagemath.blogspot.com

Post reply on HN
Print view View original

71–80 of 122 posts

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#71

Post by karlgrz »

api wrote: post #26
alexbanks wrote: post #24

Earlier quoted context omitted.

RethinkDB's niche is allowing a service to listen for changes to a dataset. As in, your Node application could listen for published events by Rethink, rather than your Node application long-polling the dataset for changes. "Real Time" data analytics can then be performed.

Its other niche is easy to deploy replication and sharding that actually works , actually scales , and does not lose data.

Concur. We've been running a small cluster for about a year and we've had the same experience thus far.

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#72

Post by dkersten »

chipotle_coyote wrote: post #70
paulsutter wrote: post #4

To understand what happened to RethinkDB, you want to contrast it with MongoDB. RethinkDB is amazing software and a real breakthrough. And the audience of deep thinking people who can really leverage the event model is ... limited. Today. MongoDB is no breakthrough, but there are legions of people who just want to plug something together with PHP or Node. Given what they're doing, it works fine. I sincerely hope Reth…

I admit I don't have a lot of experience with MongoDB, but RethinkDB was really easy to plug into Node. While we didn't support PHP officially, the PHP RethinkDB driver was maintained by RethinkDB's engineering manager, Daniel Mewes, and it looked pretty solid. When I started at RethinkDB, it was before changefeeds were a thing. I always thought the "killer feature" was, well, ReQL. It had the qualities of an ORM/DB…

In my opinion rethinkdb is super easy to get going and use. The low touch maintenance is what drew me originally.

However, learning to live without transactions takes some rethinking of the approaches taken. It's totally doable but it may not come easy if you're used to transactions in an RDBMS.

Of course in today's distributed multi service multi database world, learning to live without transactions is becoming more and more important anyway.

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#73

Post by jpgvm »

DigitalJack wrote: post #61
davidbanham wrote: post #53

Earlier quoted context omitted.

That thought fills me with dread. Rethink is well engineered and has a great developer experience. Mongo isn't and doesn't.

All these comments are starting to make me feel stupid. I picked mongo for a side project, and it's been really nice so far. I have a hard time imagining how one would use it for a complicated and/or large schema though. I assume that is just my newness to it.

Don't feel bad for not understanding how to make it work for a big project, it just doesn't. You can try make it work for a big project, but at that point you are just implementing database level code in your application.

Better to just use something like PostgreSQL. Contrary to popular belief PostgreSQL is actually a better document store than MongoDB via it's jsonb type, operators and GIST/GIN indexing.

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#74

Post by jpgvm »

meira wrote: post #55
paulsutter wrote: post #4

To understand what happened to RethinkDB, you want to contrast it with MongoDB. RethinkDB is amazing software and a real breakthrough. And the audience of deep thinking people who can really leverage the event model is ... limited. Today. MongoDB is no breakthrough, but there are legions of people who just want to plug something together with PHP or Node. Given what they're doing, it works fine. I sincerely hope Reth…

MongoDB was much more breakthrough than this little thing called rethinkdb. Hell, people even create the term NoSQL because of it. Maybe the failure of this project came from this dellusional thinking I keep reading in all posts about their failure.

MongoDB coined the term maybe, along with webscale and a whole bunch of meme.

But "NoSQL" atleast as it'd defined has been around forever, it predates SQL because everything before then was trying to work out how to store bits on disk as efficiently as possible before relational theory really found it's way into storage systems and revolutionized things by focussing on modelling.

Even putting aside that fact it was a terrible database then, it's a slightly less terrible database now and everything it can do is done better by other databases.

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#75

Post by phamilton »

gamesbrainiac wrote: post #17

I think Rethinkdb although having quite a lot of ambition and delivering on a lot of that ambition just wasn't compelling enough compared to the strides made by competitors. One example I keep referring back to is postgres. Postgres is soon going to be getting distributed capabilities, and I think it corresponds to the need in having highly available systems. Now, this was what was missing from postgres all along as…

RethinkDBs changefeeds were a lot more powerful than LISTEN/NOTIFY.

Some of it is just nicer abstractions (like limit/offset changefeeds, where you get notified if say a high score list changes). But there are a bunch of subtleties in that space that RethinkDB had to solve. Like enabling you to fetch the existing results and receive the new ones without gaps or duplicates. RethinkDB 2.0 didn't have that, but 2.1 added it. They had to iterate many times to improve the changefeed semantics, and to get LISTEN/NOTIFY to the same place you'd have to solve a lot of those same problems.

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#76

Post by lucio »

dkersten wrote: post #72
chipotle_coyote wrote: post #70

Earlier quoted context omitted.

I admit I don't have a lot of experience with MongoDB, but RethinkDB was really easy to plug into Node. While we didn't support PHP officially, the PHP RethinkDB driver was maintained by RethinkDB's engineering manager, Daniel Mewes, and it looked pretty solid. When I started at RethinkDB, it was before changefeeds were a thing. I always thought the "killer feature" was, well, ReQL. It had the qualities of an ORM/DB…

In my opinion rethinkdb is super easy to get going and use. The low touch maintenance is what drew me originally. However, learning to live without transactions takes some rethinking of the approaches taken. It's totally doable but it may not come easy if you're used to transactions in an RDBMS. Of course in today's distributed multi service multi database world, learning to live without transactions is becoming more…

if "Databases should be indistinguishable from magic" then somehow you should be able to get transactions

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#78

Post by williamstein »

phamilton wrote: post #75
gamesbrainiac wrote: post #17

I think Rethinkdb although having quite a lot of ambition and delivering on a lot of that ambition just wasn't compelling enough compared to the strides made by competitors. One example I keep referring back to is postgres. Postgres is soon going to be getting distributed capabilities, and I think it corresponds to the need in having highly available systems. Now, this was what was missing from postgres all along as…

RethinkDBs changefeeds were a lot more powerful than LISTEN/NOTIFY. Some of it is just nicer abstractions (like limit/offset changefeeds, where you get notified if say a high score list changes). But there are a bunch of subtleties in that space that RethinkDB had to solve. Like enabling you to fetch the existing results and receive the new ones without gaps or duplicates. RethinkDB 2.0 didn't have that, but 2.1 adde…

It is true that there were a huge number of subtle issues with exactly what you're describing above (involving changefeed semantics). Many of the RethinkDB bugs I reported (all of which the team fixed!) were related to these edge cases, which I hit when pushing things hard in production. RethinkDB is amazing in a way that is not interesting to an academic researcher, but matters very much for production use. I really, really hope a couple of key people realize this so that RethinkDB can have a second life as an open source project without a company home.

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#79

Post by akg_67 »

williamstein wrote: post #68
Ros2 wrote: post #67

>They were also worried about how small the total market cap >I do not think it is impossible to build an open source business (I think Wordpress is an example), and I do not need to raise VC. Is the right market here not everything like A16Z told you? The market for wordpress is crazy large with several blog platforms seeing 100+ million dollar exits and it seems no one who gets in this space fails--clearly the mark…

I don't see the market I am targeting as huge, and I agree with you that it is probably easier for Wordpress to succeed than SageMathCloud. I'm perhaps too realistic about the potential market for SageMathCloud, which was a reason why A16Z wasn't interested. If I took VC funding I might not have the option to focus on a smaller market.

Just curious, what kind of resources you have and efforts you make in the area of marketing, business development, and sales of SageMath and SMC? Have you taken any external funding for SageMath and SMC beyond grants?

The market may not be large enough and opportunity may not be as fast growing that VCs expect, still it may be possible to build a decent size business with SageMath.

Re: RethinkDB, SageMath, Andreessen-Horowitz, Basecamp and Open Source Software

#80

Post by williamstein »

gertef wrote: post #77

SageMath is an open-source alternative to Mathematica. If it goes closed source, what's the point?

The main commercial (but 100% open source) product I'm talking about in the post is SageMathCloud. To clarify, unlike SageMath, SageMathCloud is a web-based realtime collaborative environment for editing LaTeX files, writing code, using Terminals, and distributing and grading homework, which tightly integrates SageMath and Jupyter notebooks.

Regarding "If SageMath goes closed source, what's the point?" there's absolutely no possible way SageMath itself could go closed source, since the copyright is spread across over 500 contributors, and SageMath depends on many other GPL licensed libraries. In my opinion, SageMath also has much greater depth and breadth of functionality than Mathematica for many areas related to pure mathematics research, especially coding theory, number theory, algebraic combinatorics, graph theory and group theory. To a mathematics researcher such as me, the functionality of Mathematica for elliptic curves and modular forms (my area) is very basic compared to what is in SageMath.

Post reply on HN