Hacker News Reader

Comment by leephillips | Hacker News Reader

I am not an expert chef. I can cook pretty well, and my guests enjoy my food. But it doesn’t compare to what a professional chef can do.

I am not an expert abdominal surgeon. If we are in Antarctica and your appendix becomes inflamed, I will try to save you by cutting it out. But you will probably die.

I am not an expert front-end developer. I make sites for myself, for others, and have even been paid for it a few times. But I know I am basically an amateur. However, my sites are far, far better than almost all sites in the wild, created by teams of full-time, professional front-end developers. They work identically in all browsers, are fast, easy to navigate, and people say they look good. They validate, too, except for a few nonstandard htmx attributes.

There is something very strange in the WWW, when a self-taught amateur like me can make a site that is better than the New York Times’.

View on HN

Your sites are better at being tools ordinary human beings can use to their own benefit. They are (probably) terrible at generating money via ad revenue and tricking people into buying subscriptions that are then nearly impossible to cancel.

Making a site to share info and have fun is easy. Nearly anyone can do it. Making a site to actively exploit people in the most intense way possible while still being legal(ish) takes highly-trained experts.

Exactly. But I am using the measure of quality that I think is relevant. We call a chef good when we like the food, not when he finds a way to enrich a fast food chain.

Do your sites include interactive graphics and storytelling like the New York Times does? By what measure of "quality" are you setting yourself as "better" than the New York Times team?

I was circumspect in what I said and did not say. Rich Harris, the creator of Svelte, works for the Times. I use Svelte frequently, but would not have been able to create it. I don’t have his skills. And the interactive graphics that you are talking about are awesome. I don’t know how to do that, nor do I even understand how they are done some of the time. I was talking about the experience of navigating the front page and reading the articles, which is pretty bad. I am a better designer than whoever designs these pages. But I am not more skilled.

There is something very apt in your first line: A lot of sites suffer from too many cooks in the kitchen.

If every department proves its worth by claiming space on the front page, customer experience will be the last concern. That's before someone notices how giving content away is not a monetization strategy. Then add the siren song of 'only one more script' for marketing data,even if nobody has a clue about what to do with that data.

It's like Million Dollar Homepage in a sense. Every stakeholder in the business wants their stuff in there somehow. Except we're gravitating toward the million Kilabyte Homepage.

not saying that you’re wrong, but do you have the same constraints as the “crappy” websites?

there is definitely an explosion of tooling, frameworks, etc, but past some obvious things, IMHO most websites are crippled by business people (that will likely not be using the website) make all sorts of technical decisions what must go in and how it’s supposed to work (ads and tracking crap is one of the things that jumps up).

No, I don’t have all the same constraints. For example, I would walk away if any of the organizations that I work with insisted on tracking visitors. I stopped taking paying customers years ago for this and similar reasons. But if a site is bad for the reader, it’s bad. The developer has failed, even if the customer got what it wanted.

i get it and i am in the same boat as you (ie make it awesome for the reader), but unfortunately this is not what happens in the real world :(

The measure of what is better in all these companies is determined internally and generally by the business. So it ultimately serves the business and only serves the user indirectly—-if at all. Obviously you optimize for different things—-performance, reliability, simplicity perhaps. And the business has more stakeholders who want different things, more metrics, more integrations with their third-party tools, etc. That’s not to say your measure of better is wrong. It’s probably not and I probably agree with you! But it’s coming from a different place I think.

I appreciate your comment. But isn’t my conception of quality the only one that matters? If a doctor saves money for her employer by skipping some expensive test, and my health outcome is worse, her employer may be happy, but we don’t say that she is a good doctor.

here is proper doctor comparison: imagine you are a plastic surgeon and patients comes in with stack of cash and demands you make a surgery that is not medically necessary and is high risk for him - and you do it anyways and then the patient leaves happy. if you refuse, the patient would just go to the doctor next door and pay him.

This is clearly a terrible doctor. Am I wrong? Isn’t this doctor violating the Hippocratic oath?

plastic surgeons usually perform procedures that are not medically necessary, but only improve appearance to boost patient's self-esteem (boob job, facelift, butt job, etc)

You moved the goalpost. You stipulated high risk in the comment I was replying to. And who cares what plastic surgeons “usually” do? A doctor who gives boob jobs to women who are already “normal” (not to reconstruct after a mastectomy, for example) is a scourge who is exploiting society’s bad attitudes and exploiting his patients. Not a good doctor. Or are you suggesting that if a lot of people do it, that makes it good?

> I make sites for myself, for others, and have even been paid for it a few times.

Here you go. You make sites for yourself. My own stuff is fucking fast as well.

It gets slow when you have to ward of a thousand idiotic requests and implement maybe 10 of those to shut people up.

While in my own world I’d leave it as it is.

Shitty product is primarily the result of shitty culture. It’s just that FE is visible and atrocious DB calls are not.

It's kind of like saying that you as an amateur chef can out-cook the line cook at a cafeteria who prepares a thousand meals because your best-prepared meal is better than their offering. That actually may be true for that specific case. And sure, your site works great when a couple of people look at it. But what happens when you direct the entirety of the NYT's traffic to it to see how it does?

No, not at all. It is more like saying that I make better food than almost all restaurants in town. Which of course is not true, and that’s the point. Because I do make better websites than almost any I come across in the wild.

Your question about handling traffic is orthogonal to the topic of design. The answer is that any of my sites would do better, given the same server architecture, because I deliberately limit the amount that needs to be transferred for any particular page.

I think their analogy may still hold.

My mom claims her cooking is better than the restaurants (arguable) but she's not operating under the constraints of cooking at the variety, consistency and speed that my local diner does. The diner needs to be able to provide hundreds of dishes on short order including on days when the main chef is out. So maybe my mom's once-in-a-while pot of chilly is great but she couldn't run scale to run a profitable diner.

Similarly, sounds like you are hand crafting awesome websites on your own whim and schedule. That's awesome - me too - and I've done amazing handcrafted HTML/js/css for fun.

And then I go to work and manage an organization whose front-end experiences are decidedly not hand crafted and I feel great about that. Our product helps people navigate one of the most important decisions of their life and and sacrificing our ability to iterate quickly at the expense of hand crafted front-end code would be the wrong call. Like my mom's chilly it wouldn't scale.

You'd look at my work product and say "ugh what a badly crafted website" and I'll take the criticism but I wouldn't change it.

I’m not talking about sites that aren’t beautiful. I’m talking about sites that are user-hostile, annoying, and a pain to navigate. And that seems to have become most of them.

Why do you think your mom's chili wouldn't scale? Do you have any material reason to think that?

I think non scalability is the default assumption. Why do you think it WOULD?

If I stumble into the diner at 3am drink and want a bowl of chilly, guarantee positive outcome. At my mom's, not so.

Thanks, this was exactly the thrust of my analogy. I appreciate it when it lands for at least one person (though perhaps I could be a bit more precise in my writing)

You're welcome! I rely on analogies a lot and I find them helpful. Turns out many others think differently and don't find analogies an inherently helpful way to think. They get confused and get caught in details ("but chilly is spicy and that's good, but you don't want spicy code so... what?") It's just cool to see how many thinking models there are out there.

I’m an amateur builder but the work I’ve done on my house is better than most of the work I’ve seen done by various trades on my friend’s/family’s places.

Quality you get from strangers correlates with how easily the average customer can judge the work (food is easy to judge). Sometimes things are important enough to be regulated (healthcare) otherwise most markets are for lemons.

Interesting point. But as far as websites go, we all suffer from their horribleness, yet they keep getting worse.

> There is something very strange in the WWW, when a self-taught amateur like me can make a site that is better than the New York Times

Would love to see something that does everything the NYTimes does that is clearly and obviously "better."

Easy: take the front page and remove every headline and introductory paragraph that, for some reason that boggles my mind, is repeated, sometimes more than once, sometimes more than twice, on various areas of the enormous page. Now you have a page with the same information that is lighter and easier to find things on. And less stupid.

Another example: sometimes there is a stock ticker near the top of the front page, and the number of digits it displays changes as the ticks go by. But they did it wrong, and when this happens the entire layout jumps. I learned how not to make this kind of mistake near the beginning of my self-education in amateur web design.

I think the core mismatch here is that you are optimizing for experience and NYT, as a business, is optimizing for revenue. Those two things often do not overlap.

I can virtually guarantee you that NYT has spent thousands upon thousands of hours optimizing and multi-variant testing their headline display to maximize conversion. In other words, if they followed your advice, they would be leaving money on the table.

As for your stock ticker display bug, keep in mind that NYT likely employs hundreds of developers and dozens of teams each owning various pieces of the website. Bugs will make it to production, so it's a matter of prioritization. I'd be shocked if they weren't already aware of that bug. It's probably lower on some specific team's backlog than a bunch of stuff that will be more valuable for the company in terms of revenue.

You are making my point for me: “you are optimizing for experience and NYT, as a business, is optimizing for revenue.” That is simply an attempt to explain away the bad experience of reading the Times. The explanation is probably correct. But by offering explanations for why the site is bad, you are implicitly agreeing that it is bad.

I wouldn't call it "bad". I'd say it's not perfect, and rightly so, because perfection would be lower ROI than some other things their developers could be doing.

Neither your cookings or your websites have no pressure from business side, right?

Yeah it's strange. For example, Facebook has billions of dollars, thousands of developers, and they themselves created the front end framework that they are using - yet Facebook's front end is slow and glitchy.

Reply on news.ycombinator.com