Hacker News Reader

Hacker News

Comment by confluence | Hacker News Reader

Parent Full thread

confluence

Reminds me of a Charlie Munger quote:

that to get closer to the right answer, scholars should "invert, always invert."

You have a strong point there. Most top 100 websites I see are just frontends to databases.

I kid you not.

Their competitive advantage is proprietary data holding and network effects to get more data.

  Google - search + database frontend.

  Facebook - database frontend.

  YouTube - database frontend.

  Yahoo! - search + database frontend.

  Baidu.com - search + database frontend.

  Wikipedia - database frontend.

  live.com - database frontend.

  Twitter.com - database frontend.

  QQ.COM - database frontend.

  Amazon.com - database frontend.

Most of the stuff is just a front end to a database.

Their positive marginal benefit derives from holding that data, using it as leverage to get more data, and then serving it back to their users in a tight loop.

View on HN

Interesting point. However, I'll disagree with Amazon.com. Purchasing real world products is much more than just a database frontend.

If you're talking about warehousing/logistics - that is true.

However they only did it because they had to reduce costs and simplify processing. I'm pretty sure Amazon would be much happier to just dropship by making an API call to third party logistics - which is essentially an inventory database from Amazon's point of view - but they are just too big now.

I think they still do that with long tail items (don't store, but use third party logistics with an API call).

Would not having a 3rd party handle the logistics cut into their profit margins? I would assume because they do it in house and at such a large scale they save money vs hiring a 3rd party that would want their own cut raising logistic costs.

Yes. They did that at the beginning, but since they scaled up they have brought a lot of it in house.

I was addressing the fact that even though they did logistics/warehousing, I still believe them to be a database frontend (inventory/customer/credit cards), and I disagree with the above statement:

> However, I'll disagree with Amazon.com. Purchasing real world products is much more than just a database frontend.

I see the two as separate things.

Amazon.com, the website, is a database which churns out orders that manifest in the real world. But once that happens is not amazon.com the actual site any more. That has done its job of gathering data, processing it, and outputting it.

Its like if I phoned in an order. The web site is like the phone system. The phone system is not doing anything in the physical world, except passing on the data of an order.

That said, I can see a right old debate about there the lines of separation are.

Given that really the only thing a computer can do is store, manipulate and present data, what else would make sense to store that data? Aren't RAM and hard drives pretty much databases?

I was merely pointing out that there isn't anything special with most of these sites. No hate. I love them, and use most of them on a daily basis.

Just pointing out something I found highly curious.

You might as well say "Most top 100 websites I see are just /code running on a computer/"

Correct. I shared this thought because I believed that the realisation that most websites are simply database frontends was an unique, interesting and somewhat liberating thought.

I apologize if it feels like I'm stating the obvious, but it was a startling realisation to me.

I always kept each site/company in their separate boxes, when in fact they all do rather similar things, with different branding.

Yes, since the 1970s or perhaps earlier, most computer software has been "forms" (screens for a user to enter some information to be stored in a DB) and "reports" (screens that present information stored in a DB to the user). That's it. Green screens, client-server, web sites, etc are all variations on a theme.

The "social network" aspects of FB could perfectly well have been done on an IBM mainframe with 3270 terminals running IMS, 40 years ago.

I don't think that's a just argument; a database is just a means of storing information which is parsed data, the important part there is the transition from data to information, not the fact that they use a database to store the information. The reason that they are all top websites is because of their first stakes, ingenuity, pioneering, etc.

Well none of them were the first to do what they did. None of them actually pioneered anything, the ideas were in the air.

Google was an improved HITS algorithm and the infrastructure of infoseek (copied). They weren't the only search engine to do link analysis. There were others.

You think Yahoo! was the first portal? Or Netscape the first browser?

Facebook was not the first social network. MySpace/Friendster were also pretty big back then.

You can say they executed better, but at the time, my friend, it wasn't such a sure bet.

That's the way it is with innovation, and that's why companies fail so hard, and so fast. No one really knows what's going on.

You give them fundamental attribution due to the survival bias found in innovative industries.

You forget the many failures.

I'm sorry, but you mistake what I mean by pioneering; they may have not been the first in their field, but their methods were pioneering, otherwise they wouldn't be as big as they are.

Facebook wasn't the first social network, nor was MySpace, but the fact that they are the biggest now says something. The same can be said for Google; Yahoo! was massive, but their priority wasn't search, it was being a portal.

Also, failure isn't necessarily a bad thing; who are we without our mistakes?

What's a good example of a site that isn't a db frontend?

Wolfram Alpha and the search part of Google/Yahoo! - also bing.

These do more than just regurgitate data. They do useful things, such as compose summaries, search and analyse the links etc. Most of the others is simply a "SELECT * FROM TABLE" for each page.

This begs the question. Websites are basically databases connected to servers through some controller logic.

And isn't that the point, that now we're in an information age?

Reply on news.ycombinator.com