Hacker News Reader

Comment by hoodoof | Hacker News Reader

So half marks on search for Azure Cosmos DB.

If free text search is such a hard problem then you'd think that would be even more reason to start with solving that toughest of all problems. If it's a really hard problem then it will be even harder to retrofit later into some system that is already architected and built.

A light spanking for the architect. No search would have been a thorough spanking.

View on HN

The good news is that you can configure all the data to be copied over to Azure Search on a regular basis and then use search capabilities provided by Azure Search. This may sound like a complicated thing to do but the functionality provided by Azure makes it ridiculously easy to configure something like this.

Search should be a first class feature of a database not an addon. Every.Single. Time. I have ever needed to "just add on search" it has resulted in deep pain.

Nothing you say will convince me that it's just a super easy dance in the daisies to "just do X" or "just do Y" and suddenly its database search heaven.

Database application architects need to wise up that the world revolves around search - built in, not add on.

Search is not hard. Microsoft are being lazy. Which is good for me. I'm building what I thought was a Elasticsearch killer. Seems it's also a DocumentDB killer. Search is first-class with Resin: https://github.com/kreeben/resin

Reply on news.ycombinator.com