Hacker News Reader

Comment by xyproto | Hacker News Reader

After reading the article, I am left with the impression that Steve Yegge was mostly wrong.

Yes, the blog author was pretty generous. He got almost everything wrong if you're being at all critical.

The smartphone market is at least 5 years out. 3 years later iPhone comes out and defines the market. Saying it didn't get really, really popular for a year or two is misleading. If you were betting on the market being _at least_ 5 years out you would have missed that boat.

I'd say he was correct on #3, #6, #7, #8. That is 4/9, because #10 was just a joke. So he got roughly half of it right, which is not too shabby.

A flipped coin could have gotten that same fraction. ;)

A flipped coin can answer a question, but it can't ask the question. You can flip a coin to answer, "will someone invent a new way for people to socialize online?", but flipping the coin does not grant you the insight to even ask that question.

Do you not consider #1 to be at least close to accurate? Instead of XML we got JSON databases (NoSQL) which are very popular in some circles. They definitely do not surpass the number of relational database installations, but I wouldn't be terribly surprised if they stored more data in absolute terms.

No. Yegge bet on XML specifically with all its XPath, XQuery, DTD stuff. I don't believe he was thinking about anything like NoSQL or JSON.

The quoted reason is “Nobody likes to do O/R mapping; everyone just wants a solution.” From that point of view, I think it's at least arguable that he chose XML databases as a stand-in for NoSQL, which wasn't really a thing yet. The fact that many NoSQL databases use JSON, which is the successor to XML for data interchange, lends a bit more support as well IMO.

In other words, I think saying XML databases are the future in 2004 is like saying JSON document stores are the future now. I don't view either as all that correct, but there's definitely been a surge in interest.

I'd say that non-sql databases did indeed explode in popularity, so this does get the nod. However, I think the prediction may have gone a little too far. Have they surpassed relational databases?

I'm getting the impression that a lot of people are reconsidering the "nosql" approach, as well, sticking with (or reverting to) sql databases, including simple ol' mysql. This was an interesting interview that showed up on HN a few days ago.

https://medium.com/s-c-a-l-e/github-scaling-on-ruby-with-a-n...

I was never all that enthusiastic about "nosql", partly because personality wise, I tend to be reluctant about new technologies, and I liked a system based on relational algebra, all that. That said, "nosql" wasn't really meant as a NO to sql. I get the feeling that the term got away from them a bit, largely because "you-know-sometimes-there-really-are-better-solutions-than-sql-for-some-problems-and-you-might-want-to-consider-them" isn't really an option as a catch phrase.

That said, nosql absolutely did grow massively in popularity, it did move from the fringes to mainstream use, so as a prediction, yeah, I'd agree this should score a few points.

Reply on news.ycombinator.com