Hacker News Reader

Comment by iopuy | Hacker News Reader

"you want a clone for eBay/Facebook/Amazon? You realize they have teams of many thousands of people working on code bases over 15 years old right? And you think a single person, me, is going to be able to replicate that in 3 months?"

View on HN

"You can use Lisp if you want to."

"...I'll have the POC on your desk by next week."

And it will take hundreds of people over years to go from the POC to an equivalent product.

Because they'll do a costly re-write in python? :P

Because it took hundreds of good developers to find one who was experienced and was willing to work in lisp. /s

I remember one chap who wanted us to build a complex risk management application for investment banks who got most irate when we asked the details of what it should do (NB purely the required functionality, we could work out to technically build it).

Eventually we did persuade him - we should have given up early on as eventually after many many months of work the whole thing ended up in a deeply stressful court case.

Edit: After the litigation settled with anyone walking away and paying their own legal costs he approached me and wanted to be friends again..... Didn't reply to that email.

Edit2: This was 20 years ago and the litigation still makes me stressed thinking about it now :-|

Stop editing your post and give us some more details, now that you made it interesting :)

> complex risk management application for investment banks ... details of what it should do

As someone who works on risk and regulatory management software for investments banks: often they don't know exactly what they want it to do. Well, they know what they want it to do right now, vaguely, but what detail beyond the vague that they give will differ from what they'll want next time you meet with them.

Our lives would be much easier without users!

Spent a year working on an ecommerce platform with two other people. The company wanted their own custom thing, but at the same time to launch a new company through us.

Lead kept wanting to imitate Shopify. I kept saying if we try to compete with Shopify we're already dead -- we're just a small team with maybe a couple years runway; it has to be different.

Eventually the company we were working for abandoned the project and switched to Shopify, since, you know, we were building a poor copy of that anyway...

I see people on this very website claim they can probably implement an Uber clone in a weekend or two - and they're probably right, the basics of such a service are quite simple, but perfecting it and marketing it are far harder.

Is that my problem though as a freelance developer? Should I invest effort in something knowing it will fail because it can't possibly be marketed successfully against competitors?

The linked blog post is really about bad customers, and either firing them, or not accepting them in the first place. As a freelance developer it seems easy enough to say if someone's going to pay you your going rate to build something, accept their money and build what they're asking.

The problem is that in reality, clients like that often turn out to be bad customers. They don't know what they want, make all the excuses to avoid paying on time (or worse, offer "equity"/revenue share), and you end up spending all your time and energy supporting them - just one more little change and I'll pay you - and not by working on actual client projects.

Sussing out which clients are like that is up to you as a freelance developer, and it's a skill that isn't programming but is well worth your time to develop.

(Putting the funds into escrow only goes so far.)

Aargh, the memories of these frustrating meetings. I am so happy I am not developing for clients anymore.

Relevant XKCD [1]. It's the one about identifying birds.

https://xkcd.com/1425/

The proof that even XKCD can go out of date. Now (soon) we just call a google image API :).

Arguably, it's still relevant. It took about 5 years for suitable image recognition apis to come out.

Although Flickr tried to get it done in about a month: http://code.flickr.net/2014/10/20/introducing-flickr-park-or...

When the API becomes common practice the xkcd is no longer relevant, yet it was correct (at time of writing). Unless educating on the difficulties of the past of course, then it would be relevant.

I could have done it in a weekend with Mechanical Turk.

Reply on news.ycombinator.com