Comments on: Why your startup should be using MongoDB

By: MacYET

MacYET — Tue, 21 Dec 2010 05:20:00 +0000

This posting is truly nonsense. Each project requires a dedicated choice of technologies and frameworks. Recommending a noSQL database for all and everything is as stupid as using a RDBMS for all kind of applications.

By: jwhitley

jwhitley — Tue, 17 Aug 2010 23:32:00 +0000

> Why is that a bad way to convince people?

“Show me, don’t tell me.” Repeating a statement doesn’t make it any more obvious (or more true). Demonstrating that ease via examples makes it self-evident and lets the reader come to their own conclusion instead of simply being told — a much more powerful result for a technology evangelist.

By: Anonymous

Anonymous — Mon, 16 Aug 2010 00:23:00 +0000

The whole point of the article was that it makes it much easier to pivot. He mentions it in three different places in the article. That’s a fact. Why is that a bad way to convince people?

By: Sebastian Wain

Sebastian Wain — Sat, 14 Aug 2010 16:13:00 +0000

For some toy project I choose mongodb over couchdb because I can change specific json elements without the need to retrieve all the document and because the python wrappers are straightforwad.

By: Antal István Miklós

Antal István Miklós — Sat, 14 Aug 2010 15:48:00 +0000

This post failed to convince me that NoSQL stuff are easier to develop with, it just presents all this like a fact, which is certainly a bad way of convincing people.

By: filippo diotalevi

filippo diotalevi — Sat, 14 Aug 2010 13:17:00 +0000

I’ve actually come to the same conclusion (probably because we *are* using a relational DB, and migrations and changes are a royal PITA).
Relational DB have their places, and in conjunction with ORM tools it is actually possible to develop stable, maintainable and scalable applications with them.
On the other side, startups are all about fast changes, experiments, rewriting and throwing away code. In this contest relational DBs are a curse.

By: Jeffrey Jose

Jeffrey Jose — Sat, 14 Aug 2010 07:32:00 +0000

Pre-optimization is at the heart of all software evil, and it applies to data design as well. Bring SQL back into rotation after you’ve found your market, and a specific project calls for it.

Good one.

By: Justin Dearing

Justin Dearing — Fri, 13 Aug 2010 18:51:00 +0000

Startups tend to neglect hiring DBAs and Sysadmins initially. Even if they did, they usually make a lot of mistakes early on. Those that survive just make the mistakes quicker and correct them quicker. This is certainly in line with tat.

Personally I’m using mongo in two places now. In one I am storing ad-hoc registration data for one time events, so I don’t want this data sitting in my database where I’ll be tempted to come up with a data archive system. In another I need to return complex BSON objects in a situation where business rules will change rapidly, and it makes little sense to normalize all this data when it will only be queried on way. in both cases its a matter of a RDBMS not making sense.

I am currently modifying an existing set of web services that use SQL server. We now have a better idea of what the service will need to do in the future, and better realization of how much we don’t know about the future. If we could start over again, I’d certainly consider the strategy of using mongo, and switching to a RDBMS when things solidified.