4 ms·

Change feeds look great. r.http seems like a frippery, though. Not a great signal to see things like that getting added when there are much bigger fish yet to f

Change feeds look great. r.http seems like a frippery, though. Not a great signal to see things like that getting added when there are much bigger fish yet to fry.

coffeemug 12y ago
The r.http command was controversial, even internally. I had to call in a lot of favors to convince people to get it in, but here is why I was convinced it's a good idea: - It fits! It just seems to work magically well with the rest of ReQL on so many levels! JSON fits, streams fit, lazy evaluation fits, even batch prefetching fits! Everything works so wonderfully and provides such a great interactive experience, it almost would be silly not to add! - It makes a use case that's really important to me personally (adhoc analytics) an order of magnitude easier. Over the past few months I ran into a few other people that also do a lot of interactive adhoc data analysis with Rethink, which redoubled my resolve to add r.http. - It makes example datasets *much* more elegant. Reading RethinkDB docs? Just call this command to get the dataset in so you can play around with a command. (We haven't updated the examples yet, but we will). That makes the first experience with the product better, and makes the learning experience much more pleasant. So I hear what you're saying, but I respectfully disagree. If you think of it as a signal, think of it as a signal of deep passion for the product and our users! Ornamentation for the sake of ornamentation isn't what we do. (On an unrelated note, I really love your use of 'frippery' -- I'm a bit of a word nerd, and this is a really neat word!)
- davidbanham 12y ago
  Thanks for the considered response. I think I have a much better handle now on why it seemed like a good idea to you and a bad idea to me. The disconnect is that we see Rethink in different ways. You're looking at it, in this context, as Rethink + ReQL combining into a "tool for data analysis". When I look at Rethink I see "Place to put my data". My viewpoint pretty much ignores ReQL as anything more than "Means to get my data out of the thing". Which, thinking harder about it, is wrong. You guys are looking to make ReQL more than that. My viewpoint is too heavily influenced by the way I think about things like CouchDB. I normally don't go in for value judgements about people's viewpoints, but I think in this case yours has to be declared objectively correct. Given that it's your product and I'm just a guy that's been an interested observer for a while, but has never actually used it in anger. So, thanks for helping me get it. Word nerds unite!

Change feeds look great. r.http seems like a frippery, though. Not a great signal to see things like that getting added when there are much bigger fish yet to fry.

coffeemug 12y ago

The r.http command was controversial, even internally. I had to call in a lot of favors to convince people to get it in, but here is why I was convinced it's a good idea: - It fits! It just seems to work magically well with the rest of ReQL on so many levels! JSON fits, streams fit, lazy evaluation fits, even batch prefetching fits! Everything works so wonderfully and provides such a great interactive experience, it almost would be silly not to add! - It makes a use case that's really important to me personally (adhoc analytics) an order of magnitude easier. Over the past few months I ran into a few other people that also do a lot of interactive adhoc data analysis with Rethink, which redoubled my resolve to add r.http. - It makes example datasets *much* more elegant. Reading RethinkDB docs? Just call this command to get the dataset in so you can play around with a command. (We haven't updated the examples yet, but we will). That makes the first experience with the product better, and makes the learning experience much more pleasant. So I hear what you're saying, but I respectfully disagree. If you think of it as a signal, think of it as a signal of deep passion for the product and our users! Ornamentation for the sake of ornamentation isn't what we do. (On an unrelated note, I really love your use of 'frippery' -- I'm a bit of a word nerd, and this is a really neat word!)

davidbanham 12y ago
Thanks for the considered response. I think I have a much better handle now on why it seemed like a good idea to you and a bad idea to me. The disconnect is that we see Rethink in different ways. You're looking at it, in this context, as Rethink + ReQL combining into a "tool for data analysis". When I look at Rethink I see "Place to put my data". My viewpoint pretty much ignores ReQL as anything more than "Means to get my data out of the thing". Which, thinking harder about it, is wrong. You guys are looking to make ReQL more than that. My viewpoint is too heavily influenced by the way I think about things like CouchDB. I normally don't go in for value judgements about people's viewpoints, but I think in this case yours has to be declared objectively correct. Given that it's your product and I'm just a guy that's been an interested observer for a while, but has never actually used it in anger. So, thanks for helping me get it. Word nerds unite!