Post Syndicated from Bozho original https://techblog.bozho.net/blockchain-everything-looks-like-nail/
Another thing they have in common is that they are useful. All of them have some pretty good applications that are definitely worth the time and investment.
Yet another thing they have in common is that they are far from universally applicable. I’ve argued that monoliths are often still the better approach and that microservices introduce too much complexity for the average project. Big Data is something very few organizations actually have; AI/machine learning can help a wide variety of problems, but it is just a tool in a toolbox, not the solution to all problems. Single page applications are great for, yeah, applications, but most websites are still websites, not feature-rich frontends – you don’t need an SPA for every type of website. NoSQL has solved niche issues, and issues of scale that few companies have had, but nothing beats a good old relational database for the typical project out there. “The cloud” is not always where you want your software to be; and SOA just means everything (ESBs, direct integrations, even microservices, according to some). And the blockchain – it seems to be having limited success beyond cryptocurrencies.
And finally, another trait many of them share is that the hype has settled down. Only yesterday I read an article about the “death of the microservices madness”. I don’t see nearly as many new NoSQL databases as a few years ago, some of the projects that have been popular have faded. SOA and “the cloud” are already “boring”, and we’ve realized we don’t actually have big data if it fits in an Excel spreadsheet. SPAs and AI are still high in popularity, but we are getting a good understanding as a community why and when they are useful.
But it seems that nuanced reality has never stopped us from hyping a particular technology or approach. And maybe that’s okay in order to get a promising, though niche, technology, the spotlight and let it shine in the particular usecases where it fits.
But countless projects have and will suffer from our collective inability to filter through these hypes. I’d bet millions of developer hours have been wasted in trying to use the above technologies where they just didn’t fit. It’s like that scene from Idiocracy where a guy tries to fit a rectangular figure into a circular hole.
And the new one is not “the blockchain”. I won’t repeat my rant, but in summary – it doesn’t solve many of the problems companies are trying to solve with it right now just because it’s cool. Or at least it doesn’t solve them better than existing solutions. Many pilots will be carried out, many hours will be wasted in figuring out why that thing doesn’t work. A few of those projects will be a good fit and will actually bring value.
Do you need to reach multi-party consensus for the data you store? Can all stakeholder support the infrastructure to run their node(s)? Do they have the staff to administer the node(s)? Do you need to execute distributed application code on the data? Won’t it be easier to just deploy RESTful APIs and integrate the parties through that? Do you need to store all the data, or just parts of it, to guarantee data integrity?
“If you have is a hammer, everything looks like a nail” as the famous saying goes. In the software industry we repeatedly find new and cool hammers and then try to hit as many nails as we can. But only few of them are actual nails. The rest remain ugly, hard to support, “who was the idiot that wrote this” and “I wasn’t here when the decisions were made” types of projects.
I don’t have the illusion that we will calm down and skip the next hypes. Especially if adding the hyped word to your company raises your stock price. But if there’s one thing I’d like people to ask themselves when choosing a technology stack, it is “do we really need that to solve our problems?”.
If the answer is really “yes”, then great, go ahead and deploy the multi-organization permissioned blockchain, or fork Ethereum, or whatever. If not, you can still do a project a home that you can safely abandon. And if you need some pilot project to figure out whether the new piece of technology would be beneficial – go ahead and try it. But have a baseline – the fact that it somehow worked doesn’t mean it’s better than old, tested models of doing the same thing.
The post When You Have A Blockchain, Everything Looks Like a Nail appeared first on Bozho's tech blog.