Comment by version_five | Hacker News Reader

Comment by version_five | Hacker News Reader

I wonder if the future (or maybe present) of software development is a lot like construction trades. Many work for bigger firms, but there are loads of drywallers, carpenters, roofers, electricians and plumbers - though the latter two are a bit different, that are essentially independent, and get called in either directly by a customer or by a contractor to deliver part of a project.

This model works well for trades and notably is absent any real management - sub-trades are hired to deliver on an outcome, not to be "managed " and they take responsibility for the quality of their work.

Construction tradespeople I know would not be interested in demand generation or fintech software or whatever else this article is pushing. I'd expect the same from shrewd solo operators, that are focused on doing their trade well and not on getting caught up in management activities. I say this because although these sound attractive to the management class, I think selling SaaS management tools to solo operators is a tougher nut to crack than people realize

> notably is absent any real management

Except for the General Contractor taking a % of all billing to manage the entire process? Construction is what gave us Gantt charts, and large projects are incredibly well-structured and managed so that heavy, expensive equipment and material isn't delivered when it can't be used.

A general contractor was not present when I hired an engineer to plan my well system last week. Nor was one there when I had an electrician over.

If what's being 'built' is basically a copy/clone of something that a person or team has done before, sure.

Coordinating discovery and planning across disconnected "software tradespeople" isn't easy. In physical construction, it's usually pretty easy for contractor A to indicate that he is blocked by contractor B. "I can't drywall until the electrics are in". People wouldn't demand that the drywall be put in anyway just so they could hit some arbitrary "velocity" goal, for example. This is much harder to decipher and work around in software pieces, especially by disconnected vendors and contractors.

It is already like that for companies whose main business isn't selling software.

Consulting companies, or freelancers, get hired to work on tickets, or do small updates to existing software that is driving business processes, factories, .....

They usually don't have any IT in-house beside a couple of project managers that take care of all external contractors and what they are supposed to do.

It is also quite common in the games industry since it exists, hence why 3D APIs are less relevant as many in FOSS comunity think, there are contractors whose main business is porting titles and they still take less than having bigger teams in-house.

The metaphor I use is movie production because it's a creative endeavor but yes, totally cosign.

What’s stopping this model from thriving now? And how might it be realised?

Unlikely.

The interface between two pieces of construction work, done by two different kinds of contractors is, generally speaking, far less variable than the interface between two different pieces of software, or a piece of software and a piece of hardware, or a piece of software, and a business process.

It's not uncommon for software to discover that the interface point it's supposed to work with is the digital equivalent of a plumber laying down pipes made out of sugar, instead of PVC.

have you ever worked in contruction or even fixed stuff around your house?

:-) its crazy how non-fitting stuff is. for nearly every slightly different job you can better expect to buy new tools or find some other hack around.

the difference IMO is that people realize that once something is up, the cost of change is HUGE. In software, (non programmer) people somehow expect that this is not so and everything always can be rewritten differently.

Reply on news.ycombinator.com