Just some warnings against mongo being a magic bullet: > P.S. Does anyone really estimated how much will be losses in case of some > (0.00...X) transactions lost due to lack of transactional support in mongo? > I mean in stores like amazon - it's not a question, losses are huge, but in > case of an ordinary e-shop? Are they really so big? One developer cost about > 100 / per-year, so, mongo allows to build product faster and thus save cost > on development time (although, it's also subject to question). The question > is - what's higher - cost of development or lost transactions.
I think there are reasons why mongo is useful, but saving on developer costs is not one of them. When I hear people argue this position, they usually don't know much about mongo (and a lot of times aren't big on databases in general) and their hope is that they can build a web application without having to learn how their db works or how to administer it (or rather, learning mongo is much easier than learning mysql). I'm not saying people are dumb for hoping this will work, but I do believe it's incorrect and is based more on PR than reality. Having used both, my opinion is that all db's are complicated and it takes a fair amount of knowledge and effort to not screw it up when the stakes are high. If you want to speed up development (regardless of db), you can outsource the db work for a while and get consultants to train you and help you set it up and use it properly. > though Mongo is not ACID . however, when i raised that question to 10gen > during a training, one of the solutions they suggest is to do it in code with > two handshakes for committing data. not pretty but dorable. This is total marketing FUD, Murvin I am glad you didn't buy it. Any form of saying "just implement transactions yourself in your application" is pretty damning - that's a *huge* burden the db vendor has just foisted onto your dev team. You are not going to be saving any developer cycles if you have to do that. I'm not saying "don't use mongo ever." Just don't use mongo because you think it will be easier than mysql, especially if you need transactions (which you will if you are doing commerce). And if you're building e-commerce and handling transactions and money, you should have staff dedicated to db operations and security very early on in your startup. You can't measure how badly things will go if going got messed up, it's impossible to make a judgement like that ahead of time. It's an important enough part of the system that things will move faster if you have someone handling it full time. Again, not anti-mongo, but I see a lot of dev's pinning their hopes on it being a magic pony (and a lot of marketing encouraging this), and it make me want to slap a big warning label on it. Maybe postgres + elastic search is what you are looking for? It's a pretty winning combo if you're building a store. Ted -- Job Board: http://jobs.nodejs.org/ Posting guidelines: https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines You received this message because you are subscribed to the Google Groups "nodejs" group. To post to this group, send email to nodejs@googlegroups.com To unsubscribe from this group, send email to nodejs+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/nodejs?hl=en?hl=en
