Your customers don't care about JS
Vložit
- čas přidán 4. 06. 2024
- Your customers do not care about the technology behind your app. What programming language, frameworks, libraries, or whatever tooling you're using, they don't care. They care about it providing them value and that it has certain characteristics. They want it to work.
🔗 EventStoreDB
eventsto.re/codeopinion
🔔 Subscribe: / @codeopinion
💥 Join this channel to get access to a private Discord Server and any source code in my videos.
🔥 Join via Patreon
/ codeopinion
✔️ Join via CZcams
/ @codeopinion
📝 Blog: codeopinion.com
👋 Twitter: / codeopinion
✨ LinkedIn: / dcomartin
📧 Weekly Updates: mailchi.mp/63c7a0b3ff38/codeo... - Věda a technologie
yuck, I'm glad I don't use Twitter, just shows how the majority is there for the hot takes and dunking on the others.
👍👏
It can be great and terrible at the same time.
I don't know what's worse, slow dialog or loading spinner before the popup's appearance
Also wanted to mention that but forgot! Thanks for the comment
I completely agree! Your customers don’t really see most of the tech behind, and even things like waiting for popups or loading times are not as important as the value the service provided.
Delivering 100% pristine products is a fairy tale made by folks, who never shipped things to production, but have 100% coverage on their hello worlds.
I feel a point people are missing in these A vs B debates is that even if the answer is grey, there will still be a more optimal choice.
In this case the hotwire people maybe didn't see the issue because the system lends itself obscured it.
Every tool has characteristics and it is our job to know our tools and maintain our skill
This was a mature take on this. I was watching Theo giggle making fun of this app, and comparing them to Google. He make some fair points but there's a an immaturity there that needs to be addressed. Personally I was pressured into building a very large project solo and when I was done I personally thought what I made was garbage, but the customer was so happy with the result. You're right, done is better than none. There are so many factors into getting from A to Z and sometimes we have to cut C and D out in order to finalize something. If we have the time, we can go back and improve on it. But you need a thing to exist first in order to improve on it.
The only coding channel with 99% talk and 1% code snippets from twitter
It's a difficult task but someone has to do it.
I learnt this lesson about 15 yrs ago when a sales guy sat in a tech meeting (we trying to choose tech's for a project) with us and he stood up in a room of 30 and said... "I gotta break it to you guys - customers do not give a sh1t about code"... And when we let it sink in, we realised he was 100% right,,
That's why some of the best developers I know are knowledgeable in both tech and business.
@@CodeOpinion I got a raise and promoted 4 times in 36 months starting from a junior developer to lead developer, largely because I understood business and psychology. I have a business administration / marketing degree.
Many professionals only care about introducing technologies that enhance their experience to secure their next job, leaving the problems for the next person to solve. Companies without individuals to prevent this and that focus solely on results become hostages to this behavior.
I call it CV driven design. Or for our American friends, Resume driven design.
Ya, I wanted to mention this in here but totally forgot. Thanks for the comment.
Sadly this can also be driven by the upper management. Implementing microservices bcos it sounds cool. Their core business mainly is just B2B, where it's overkill
I see this at times myself. Usually when applying to a new job, if I see the developer must use a vast amount of different tech, this is what comes to mind. 3 different languages for the back end, 2 for the front end, microservices, microservices using containers, etc... At this point they must be checking boxes for resume (CV) buzz words. Or they are managing different projects - each using a different tech stack. Either case sounds like a headache I don't need. These types of job postings I avoid.
excalidraw, eraser, miro, canva, and figma devs 👀
As always, you hit the nail on the head Derek. It seems like common sense to me, but worth repeating to sooth the mind. Don’t dunk on other people, products, techniques without knowing their context. Everyone brings their own truth.
If you want to avoid this sort of problem going live, in your chrome developer tools, under the "Network" tab, turn on throttling.
Great video, man!
6:15 That's golden wisdom right there. I can absolutely relate, I'm in that exact situation now. FTE -> Half a decade of contracting + side gigs -> Back to FTE. And my perspective has absolutely widened!
Great take, really enjoyed this video
As a customer, I’m care about the page, if it load more than my life time or malfunction or privately collecting my personal information without my consent , i’m out so basically the customers who know about tech also are considered as a customer ? Am I right ? They may not care at first, but for the long run they do care and do re-search about your business practice
Based on my personal experience, the majority of encounters I have are with JavaScript developers who seem to live in their own JavaScript framework bubble. They believe that everything should be JSON, client-side rendering is a must, and all solutions must be JavaScript using React JS.
As someone that doesn't like the JavaScript breeds and their offsprings i can't agree with that, SSR is not magic, fun fact i actually got fired 12 years ago because i did it on a company project, back then it was a bit different because we didn't have the same scaling architecture and now looking at my decision i would probably have fired myself as well.
There are real issues that arise when we go this route, for example, having customers on low bandwidth connection, yes you assume that modern infrastructure is great and it works in the city but there are many places across the world where an SSR app would be bad choice, take for example logistics sector, they often deliver good to places that have no coverage at all, or even if there is some it's limited bandwidth. Your application experience regardless of which cool kid language you decided to use will still be horrible compared to a normal html/css/js version of your app paired with a good backend language such as golang for example, as even if there are bandwidth issues in that case, they will only be present when submitting forms or loading the data on the form, not rendering the page, and a good frontend developer can actually mask it so well that it's practically not an issue.
Where i personally see SSR useful is company internal tools, something that you know will most likely be used internally or only from an office connection that is stable, otherwise it's a recipe for disaster.
I do think the prevalence of such issues arises from the way people get their information these days, a lot of tech youtube tutorial channels became influencer channels because it's easy to make money talking about nonsense rather than actually teaching people, gains more retention therefore it's pointless to educate when you can be making bank, and this unfortunately does have a huge negative impact on the industry as most people these days get their information from these people that heavily push something that they either knowingly or not out of their own arrogance promote for the sake of their growth on the platform.
Also JSON might not be the best medium to transfer data, but there are lot of business decisions that go into something before it becomes mainstream and widely adopted, transferring binary packets will always be better than anything, because data can be represented and optimized much more efficiently, but then you run into the issue with costs, now you need to have a really expensive frontend team, and increase the amount of work for the backend team, because they have to work on this trivial task together to create a packet structure that both the server and the client can follow. You just increased the bills that the company has each month, as well as the complexity of the project for maybe saving few milliseconds of response time, you're doing a pet clinic web platform, the load of your userbase it 100 people on a great day concurrently, JSON and Packets will have the same performance problem is with JSON you can probably cut the frontend team costs significantly, reduce the workload of your backend team as they can just use JSON without thinking about any packet structures in place and a way to optimize the serialization pipelines.
Moral of this article is, some things just work for a reason, and we all should do better to learn and explore rather than taking our source of knowledge to at best what is people with questionable proffesional experience, just because they worked at Netflix, Google, or Other big name corpo that used to have every moron that managed to breath before suffocating for longer than 5 minutes 3 years ago.
...and conveniently ignore the fact that React has builtin hooks for everything except for fetching data from the server.
Oh, you just wanted to fetch a table and display it? Enjoy your useEffect ratnests with fun race conditions when the user clicks the button twice. And of course the json representation of the table ends up being larger than a HTML table, since html tables at least have the field names in the header and then just values + separators
When all you have is a hammer, everything looks like a nail.
@@BosonCollider React can render on the server. Don't spread misinformation.
Your last point is by far the most important, thank you for making it. The best software speeds up a workflow, nobody cares about about every little nuance unless it is egregiously slow or adds an extra click.
Yes always know what you are doing, but at the end of the day ship something that works for the customer.
There are extremist on both sides of the argument (Backend group golang, c#, php, ruby) and (frontend group, next,nest, insert 1m js crap here) so taking conclusions from either side is bad idea. For me there is nothing wrong with using javascript on the client side as long as that client side is something in the web browser, it's fine it works and there is a reason why it got popular.
A lot of people get stuck on trivial stuff, just get two captains, one to run the frontend, one for the backend, drink your coffee and don't make your life complicated.
how I see it the "needing javascript" isn't about javascript being somehow superior, it's about javascript being the only client-side programming language on the web and webapps requiring client-side logic to have a good UX.
As you pointed out the popup problem is solved by preloading, but how do you achieve that? It can only be done with javascript.
great conclusion Derek. I'm just getting more and more passionate about the top level goal of our job which is modeling the business value... And if so called garbage is enough to model the reality, why do we need to fight over some technocratic-religious patterns? They were all invented to solve some particular problems within specific context after all...
Honestly, 3 years of Hey everyday and it is a slow, high churn product that hasn't seen core improvements since launch but what has been improved is their bottom line, launching new products, optimising cloud costs. It's a great example why you would care about client JS or whatever gets the job done but server side apps just doesn't. A non-dev would say it's slow and they don't like it but you can convert that to say customer care about JS if you add nuance so yes customers doesn't know what words to use but they do care about JS
Wait so the whole thing here is that if you simulate a slow connection then the app is slow because server responses are slow? That’s kind of a bizarre take given that most of the time that’s not going to be an issue. Even if the user does have a slow connection, the app will need to make some a request to the server at some point and the user will most likely experience some loading time or potentially some other type of negative experience (like seeing stale data but the client not making it obvious that loading is taking longer than usual). A slow connection is going to produce a poor user experience regardless of how much JS you use - you might be able to smooth that over with some client side logic, but is that the most important thing if 99% of the time your users have a stable fast internet connection? Which can reasonably be expected for the businesses Basecamp and Hey serve.
Tech Twitter is weird sometimes.
For sure developers care about the app performance but that's not a correct metric, the end user want a good product and being super performant is just a plus, otherwise we will be writing all ours apps on C, C++ or Rust.
I really dislike those frameworks and language wars.
It's poorly worded but I read Ryans comment as not actually personal and more the industry at large. I've been joking for a while that React devs will do literally anything but learn web standards. It often feels like web development is a competition to stack on as many complexity demons as possible that sink up extra time and money without providing customer value and getting into the industry you follow because it's how things are done.
It feels like a big part of HTMX, Svelte and Remix is calling out how batshit insane the industry at large has gotten.
I took his comment the same way. The industry as a whole.
the web standard that requires you to do 'display: flex; justify-content: center; align-items: center' to center something? I take these twitter tech guys with a pinch of salt. It's just outrage for outrage's sake. Nobody calls out how batshit insane the web standard itself is. Ask any native devs or game devs. They never are screamed at for incompatibility issues or some aria shit whereas web devs are expected to make their site accessible to million different versions of browsers. Polyfill this, webkit that all while you are expected to implement some designer's complex figma file overnight.
If your GETs and POSTs dont return at a blink of an eye for the user, you are doing it wrong. Get off that shitty cloud hosting.
I think there is definitely a problem where people just pick up what technologies are new and shiny - rather than researching what is a best minimal fit for the job. This results in tons of overengineering which means bottlenecks and generally badly architected applications. Really, as you said, if we are just doing some straighforward CRUD stuff for example... it 'should' return at the blink of an eye! Sometimes people just need to host a small monolith on a server and you are done... but instead they go and use tons of cloud services and badly architected microservice architecture right from the get go making a bloated slow mess even though you only have 10 users.
nope 😂😂 , vdom very slow is old days . But re render whole dom more slow . 100 to 500 ms is acceptable by end user but refreshing race is another problem speed
Most of my large customers have an enterprise architecture group in their IT department. They DO care what technologies are being used.
Most people worry about things they probably don't understand, suggesting solutions to problems they don't have, and creating issues they will never comprehend. This happens everywhere!
Right but why do they care? My guess is because they are trying to justify their existence. "You can't use blah, blah because we don't like it". The proof is in the pudding. I could care less if you use some new wizbang language as long as you can prove to me it meets my needs I'm good.
Welcome to the world of professional programming. When your being paid and delivery means a paycheck or keeping a contract then how an app is a built is a matter of delivering the best version possible in the time and budget allowed. So the holy wars over software design are only as valuable as how it can guide you to the best possible delivery within the constraints applied. Every other version of programming is educational or recreational.
It's ridiculous how many developers things that they job is to write perfect code. It's not. Your job is to write good enough code. Sure if you get time to refactor and do stuff in most efficient way possible go for it. But if you have to do something quick and you need to take few shortcuts to achieve your goal - just go for it.
Obviously there is much more into it, because "good enough" means different thing when you expecting few internal users than when you expecting thousands of external users, but concept remains valid.
It's ridiculous how many developers don't realize how bad code slows down productivity and cost businesses money. I'm not saying that code has to be perfect or that customers have to care about it. But devs do have to care about code quality. Software engineer seems to be the only profession that doesn't care about the quality of their work.
@@JiggyJones0 LOL. There so many professions that care very little about quality of their work. Construction workers, much of customer service etc. I would say it's totally opposite, developers are the one of very few professions where some individuals care way too much about quality instead of focusing primary on business goal. Obviously there is probably even more developers that doesn't care at all, but those are people that you want out of your company, while perfectionist you can try to coach to be more pragmatic.
Apart from that obviously keeping code maintainable is part of definition of "good enough" in my case, but maybe that wasn't clear.
@@jz5153 You can't convince someone in tech how little people care about the quality of their work. Simply because other things don't 'break' as often which is a given when you are dealing with a million times fewer variables.
Obviously 👍
Programmers still mad someone already programmed the browser
Someone stole this guys upper lips
There's more than one upper lip?
finally someone said it
There were decent replies to some of those tweets I showed that were also trying to be reasonable and put some nuance to it.
Well. This garbage you've made, you'll need to support. And Idk about you guys, but I love myself enough to refuse to spend my time and health on developing garbage.
Been telling this to all my dev friends for ages))
Business guys want the “Buy” button working as soon as possible and don’t want to listen to your lame ui components, and architectures😅 and how you need more time to fix an unfixable “module not found” error
Every app needs thought put into it using common sense first and foremost
Well from what I can tell, common sense isn't common because everyone seemingly has a different idea of what it is.
The Hey team have already fix this slow pop up… then the topic is closed
Big thanks to Event Store for sponsoring Derek Comartin that provides very valuable content on this channel.
This is 100% WRONG! Our customers and partners care a great deal about the technologies and platforms we are using. We deal with many fortune 500 companies and they want to know that their data is safe with us. They want to know that our software and the technologies we are using still have support for bug fixes and security patches. They want to know about our networks and its security. They want to know about what security appliances we have. It's true that not all of these companies care to the same level but they all care. Many ask about our SDLC. They ask about our backup process and what technologies we us to do that.
Even the US Gov and NSA specifically have an office list of "memory" approved languages. Maybe some peoples customers don't care but many do. And would dare say the larger the contract the more they are are going to care and the more they are going to prob your company and its practices. Maybe companies have security agreements with their customers and so for you to do business with them you have to adhere to those agreements too.
Just saying....
Yeah and drupal is still the number 1 stack for GOV and Military sites. PHP NOT JS :)
Of course there are some customers that care. Given the context you're in, it might be the majority. In a lot of others, it is not at all. I thought I was pretty clear in the video that context matters.
@@CodeOpinion I thought you were clear, but I did have to do more than just reading the title 😊
Utter nonsense. You've obviously never worked on a wide range of real projects.
🎯💯