Hacker News Reader

Comment by DenisM | Hacker News Reader

The answer to "buid or buy" is more nuanced than "always build".

Yes: the answer is that building almost always turns out to have unexpected advantages.

And in my experience buying often turns out to have hidden disadvantages. The biggest of which seems to me to be that whenever we've tried to buy (both off the shelf and contracted work), it always turned out to be as much trouble and effort as just taking the time to build the damn thing ourselves.

This is silly. You did not build your own laptop or your own accounting system, operating system or your own relational database. There is a crossover point somewhere where building makes more sense than buying, but finding that point is a delicate task for each business, a lot more complex than "yeah, just build it whatever it is".

I didn't say the advantages always outweigh the disadvantages, just that there are usually more advantages that you'd expect.

I see. In that case why did you label your post as "answer" to "build or buy" question?

Because that is the more nuanced answer to "build or buy?" Of course you don't always want to build. But in borderline cases, err on the side of building, because there are often hidden advantages to building.

Are we clear now? Are you convinced I'm not advocating making your own electric power generation stations?

Yep, it's clear what you meant now.

I think this was a terminology misunderstanding - I don't see it as "the answer". "The answer" in my book is specific for each particular situation and your guideline is important to take into account for each such situation. I'm splitting hair now, but only to illustrate the source of misunderstanding.

To advance the state of the art it would help if you created a list of cases where "build" produced unexpected advantage and a list of cases where it backfired. At least it would work better for me to calibrate properly.

EDIT: It just occured to me that you answer to my answer is what curried function is to a function result. Throw in lazy evaluation and the line gets blurry. :-)

for simple things, always build is better if one is less dependent to time and money

actually, scratch that, it's even better when time and money are tight

in my case, i bought baking, washing soda and lye for my hygiene. 5kg each, each kg lasts for a year. extremely cheap (like $1/kg). only one time transaction.

i no longer go to soap, toothpaste, detergent and drano section when i go to grocery. suppose that takes 5 mins for searching, picking up, queueing and paying for each hygiene run while in grocery, it easily adds up to 52*5 = 260 mins time not wasted per year.

Nomination for bizarro post of the week :)

http://en.wikipedia.org/wiki/Beowulf_(computing) comes to mind :)

Don't forget to buld your own CPU and your own RAM chips while you're at it.

If you don't have money (or you're a cheapskate), it simplifies to 'always build'.

Reply on news.ycombinator.com