3 ms·

Everyone is going to miss the point here and go down rabbit holes about various technical bits and that is exactly the point: developers spend lots of time deba

Everyone is going to miss the point here and go down rabbit holes about various technical bits and that is exactly the point: developers spend lots of time debating everything The limited resources really doesn't matter. The tech choices really don't matter. Pick a prominent MVC based web framework, a RDBMS, some front-end libraries and go build the thing. Lean towards picking stuff the team has experience using. Then if the web app gets heavier traffic, horizontally scale the web tier and vertically scale the DB to your companies budget. If that amount of scaling doesn't suffice - do the research as you approach that level of traffic not when you have zero users. When it comes down to standard web apps there are known patterns and known solutions out there to solve the majority of the problems. Reaching for a new/shiny thing to use is fun, learning new development techniques is fun but its rare either of those is going to cause a fundamental shift in solving problems with web apps. We have been using (abusing?) HTTP/HTML/JS/CSS for 25+ years. For small web apps I bet people could solve things with simple PERL/PHP cgi-bin types of scripts today with current versions of HTML/JS/CSS, albeit more securely than in 1999. To use a building a house analogy, sometimes its like all the carpenters are standing around for days talking about what tool is better to frame the house, one uses less nails, one is faster, one is a nail with wood glue that will expand at the joint and provide less squeaky floors (I made that up). None of it really matters, you're being paid to frame a house, not do research - just build the thing.

julianlam 4y ago
> To use a building a house analogy, sometimes its like all the carpenters are standing around for days talking about what tool is better to frame the house The way I see it, modern web development isn't about debating which wood screws to use or which lengths of wood... it's arguing about which pre-fab construction to use (a framework), and whether it's acceptable that if you want to build a KitchenObject, you have to start off with a RoomObject and add PlumbingFixtures, but you end up with a ClosetChild in the kitchen because it was inherited with the RoomObject. Meanwhile I'm just here building with 2x4s and wood screws, watching the debate rage on :)
tinalumfoil 4y ago
> Pick a prominent MVC based web framework, a RDBMS, some front-end libraries and go build the thing. It's really not that easy. At very least the framework you use needs to be well thought-out so you don't run into roadblocks later, and your application may very well lend itself to certain technologies that won't be obvious if you don't spend some time thinking things through and understanding tradeoffs. When constructing a building you don't just choose a material, a layout, etc and get to work. There's significant effort in understanding the environment you're building in so you can inform your choices. Buildings have collapsed because the soil conditions weren't well enough understood.
smaudet 4y ago
Hmm. Antipatterns do matter though - red tape abounds and is rarely conducive to good apps. Don't, e.g., bother to write comments for all your functions, modularize everything, or insist that every line of code must be tested. Not that those things can't be good in small amounts, 95% of modern web dev is busy nonsense like this, and it shows. Tech decisions are another source of waffling, to be sure. There has been more than one rewrite of a product to a worse product, to a product being again rewritten, or practically rewritten, because the rewrite was bad. Yeah some of it is management sticking its nose where it doesn't belong, but yeah, I think the point is also that devs argue about mostly what doesn't matter, instead of focusing on what does matter.

julianlam 4y ago

> To use a building a house analogy, sometimes its like all the carpenters are standing around for days talking about what tool is better to frame the house The way I see it, modern web development isn't about debating which wood screws to use or which lengths of wood... it's arguing about which pre-fab construction to use (a framework), and whether it's acceptable that if you want to build a KitchenObject, you have to start off with a RoomObject and add PlumbingFixtures, but you end up with a ClosetChild in the kitchen because it was inherited with the RoomObject. Meanwhile I'm just here building with 2x4s and wood screws, watching the debate rage on :)

tinalumfoil 4y ago

> Pick a prominent MVC based web framework, a RDBMS, some front-end libraries and go build the thing. It's really not that easy. At very least the framework you use needs to be well thought-out so you don't run into roadblocks later, and your application may very well lend itself to certain technologies that won't be obvious if you don't spend some time thinking things through and understanding tradeoffs. When constructing a building you don't just choose a material, a layout, etc and get to work. There's significant effort in understanding the environment you're building in so you can inform your choices. Buildings have collapsed because the soil conditions weren't well enough understood.

smaudet 4y ago

Hmm. Antipatterns do matter though - red tape abounds and is rarely conducive to good apps. Don't, e.g., bother to write comments for all your functions, modularize everything, or insist that every line of code must be tested. Not that those things can't be good in small amounts, 95% of modern web dev is busy nonsense like this, and it shows. Tech decisions are another source of waffling, to be sure. There has been more than one rewrite of a product to a worse product, to a product being again rewritten, or practically rewritten, because the rewrite was bad. Yeah some of it is management sticking its nose where it doesn't belong, but yeah, I think the point is also that devs argue about mostly what doesn't matter, instead of focusing on what does matter.