[kbenson]

Really it's very easy to build your own web application framework these days, existing web application frameworks look an easy solution for beginning your project but they become a burden gradually as your project grows.

Really it's very easy to build your own database software these days, database servers look an easy solution for beginning your project but they become a burden gradually as your project grows.

Really it's very easy to build your own car these days, pre-made cars look an easy solution for beginning your project but they become a burden gradually as your project grows.

...

All of these cases may be true if you are the 0.01% that wants to do it to learn or have some very special business need that calls for it. For the other 99.99% of the time, it just keeps you from focusing on the stuff you actually need or want to accomplish.

[tengbretson]

These comparisons are incredibly dishonest.

[kbenson]

I don't think so, and depending on your familiarity with the topic, your opinion might be different too.

Creating a web application framework that handles your needs can be incredibly easy. I can, and have multiple time, done so in Perl for a quick project in the distant past. It takes about 20-30 lines or so on top of loading the modules you would probably already want available for a system like that. Pull in HTTP::Daemon, pull in the HTTP modules, tie them together with a little regex foo for route parsing, done. In the very simple case, it's easier to do this than pull in a framework. In the very complex case where you need intricate control over odd parts of the process that frameworks don't usually give you, it's easier if you had written your own from scratch. In all the other cases, which I feel I was being generous in saying was only 99.99% of the time, you're better off having offloaded the concern to someone else.

Creating a database is also very easy. In the simplest case, you choose a record structure and write to disk opportunistically or on every change. Again, for the very simplest case, that's actually easier than even pulling in SQLite. In the very complex case, you can do interesting preprocessing and specialized storage techniques to store your data in ways that aren't amenable to anything except for a custom format.[1] Again, the vast majority of the time, you just use SQLite, Redis, or a RDBMS and it does a passable job at dealing with all your data integrity and warehousing needs. If you're a company trying to manipulate a dataset in a heretofore unseen way (either much faster or much more detailed that before), you might actually need to do it from scratch.

Creating a car is also fairly easy, if you have the prior experience (about as much or less as expected above). Like above, you're largely throwing existing parts together and the hard work is knowing what goes where and shaving off the pointy bits that prevent it from fitting really well. If you need to race something (because why else do you create a car?), in the simplest case you might just grab some cheap broken down beater and put a little work into it where needed and you have something that goes in a circle and you won't care too much if it barely finishes. If you have the money, you can buy a pre-built race-ready car and compete as am amateur. If you're an F1 team, you build the whole thing from the ground up to exact specifications.

This is, of course, all presupposing that the people involved actually have the knowledge and ability to do the really simple or advanced things. Often, they don't, or at least don't have enough to make it even feasible. For example, I know how to read and write CSS, but I'm not very good at it. Similarly, I know how to write HTML and apply CSS fairly in an amateur manner. I'm not a UI or UX designer. Does it really behoove me to learn the intricacies of design layout instead of using Bootstrap? I either don't care about how it looks at all, in which case I'll make it text/plain or at the most a couple of header tags, or I care enough that to make it somewhat passable visually I'll use Bootstrap. In what situation am I better off using CSS? In what way does that apply to more than 0.01% of the people out there?

1: https://news.ycombinator.com/item?id=15350115

[cocktailpeanuts]

Come back after you've built a database system used by tons of users and a car that someone is actually driving.

No one will take your argument seriously when you say ignorant stuff like "Creating a database is also very easy", or "Creating a car is also fairly easy".

No matter how you spin it, it is FAR from easy to build any of those, and the only people who think it's easy are those who don't have real experience.

[kbenson]

> No one will take your argument seriously when you say ignorant stuff like "Creating a database is also very easy", or "Creating a car is also fairly easy".

Holy shit, did nobody read the last sentence of the original comment? I'll spell it out. The first three things were meant to be similar statements to the comment I was replying to, and the final sentence was explaining how they, plus the version in the comment I was replying to, were bullshit.

I then went into excruciating detail in the next comment on why, for each specific case, and capped it with a paragraph summarizing my point.

> No matter how you spin it, it is FAR from easy to build any of those, and the only people who think it's easy are those who don't have real experience.

Which is fully in line with exactly what I was saying.

[DougWebb]

It's much easier to build something that's suitable for your one specific use case than it is to build a general-purpose framework. Database systems, cars, and Bootstrap are all general-purpose. But if you need to store a collection of name/value pairs, carry a load from fixed-point-A to fixed-point-B, or create a web-page with a specific and custom style, building something from scratch can be the right way to go, and it can be easier than forcing a general-purpose solution to fit what you need.

[always_good]

But we're not talking about something with tons of users.

We're talking about a bespoke just-for-my-project sort of thing. That you think it's not even worth it unless you can get tons of users is part of the point. It's just a distraction just like your bespoke CSS framework.

[cocktailpeanuts]

> Really it's very easy to build your own database software these days

> Really it's very easy to build your own car these days

Wow really? Hold my beer while I go build some car and a database! You learn something new every day!

[kbenson]

Maybe you should read the last sentence again. I think perhaps you missed something.

[cocktailpeanuts]

hah! sorry about that, i indeed did!

That said, if you have to explain your joke, you've failed :) It was too long!

[hmexx]

I actually did not even need to read the last sentence to get that he was being sarcastic

[cocktailpeanuts]

good for you!