Possibly a dumb question,

Is there any limit to how much data you should keep in memory in a browser, like say you cache all json responses in a react app in a redux store? For example, say I have an archive of a forum with millions of posts, if theoretically someone was to visit all the routes without refreshing, then they may end up with a ton of data in the reducer. I guess its hard to use too much memory if it's all just text but is this ever a concern?

Our react/css contractor is kind of flaming out. Which is funny because this is the least demanding job ever as far as timeframes. So if anyone in LA is interested - the pay is pretty good. It's 6 months contract to maybe hire (they offered me and another person). Then you get sweet university benefits. The hourly was the highest I came across when interviewing, and the full time offer they've floated to me is surprisingly competitive with the startup jobs I interviewed at.

Downsides: workplace is kinda quiet and dull, no snacks (upside for me), no wfh (yet) but flexible hours.

If you hate open offices, this is the place for you - beige cubicles. I have 12 cabinets and shelves and my desk came pre-stocked with $20 worth of office supplies - 99% of which I've never used. It's like a throwback to the 80s.

PM for details.

lostmypw: not a dumb question. Definitely don't go over 5000 rows. 1000 is a good standard limit to stick to before server side pagination. I posed the same question to Cory House (twitter wonk) when we hired him for training and he said 1000.

suzzer99: why not just say how much it pays?

If he's not the hiring manager he can't know, he can know how much they offered him and may not want to share that info with internet incase the opportunity for another wil-type "who makes more" bet presents itself.

Oh God, I think that’s still my favourite 2+2 thread derail ever. I’m going to reread it now.

Yeah I have no idea how much it pays. I just know what they offered me. And my lit negotiation skilzzz to get $5 more/hour.

It's not even open yet. But this guy has 5 weeks left and I just don't see it working out for him. He did a no-show, no-call on Monday. I think he was just hoping no one noticed. I had to badger him for a week to check in code that he's been working on for a month.

Graphql seems like a solution just for internal apis, does that sound accurate? It doesn't seem like something you could or would want to use for an API designed to be consumed by third parties

I didn't even think about public APIs but yeah I think if you had to do either for a public API REST is probably more appropriate since the bar is lower. With that said, it looks like a Github's has a GraphQL-based public API?

Today I had to tell a senior back end guy that the browser freezing after an api request returns a 32mb, 150k row response was not, in fact, a "UI defect".

We had to have similar conversations with the boss at the side job how a browser storing and assembling encrypted chunks, the decrypting a 2GB document was probably not something that should be in the MVP.

I'm currently having to explain to someone that having a system that making several hundred thousand requests/day for something that's easily cacheable on the client side is not a good idea, and it's an even worse idea when there are thousands of clients and we are hoping to expand to tens or even hundreds of thousands of clients in the next year.

ETA:
(The bug ticket that was opened was that the client was getting throttled, which, yeah, that's how throttling works)

What's their argument against caching? Possibly stale data?

Are the results always the same or the just generally the same and it doesn't matter much if they are a little out of date?

I haven't really asked aside from, I guess "we don't want to?" My suggestion is that they actually just essentially keep a local clone of the database, updating at regular intervals. The database has about 100k records, maybe 1-2kb each, that change very rarely, so they could do a daily or hourly update that gives back a record or two.

It's unusual for an old record to change, but it does happen from time to time.

They're also using an endpoint that annotates the data with a lot of stuff they don't need (the endpoint is intended for use with the UI, not really for external consumption) so they're also causing my side to make a bunch of queries for data they immediately throw away.

We'll get it worked out eventually.

The "client" in this case isn't a web browser, to be clear, it's an enterprise security product. So keeping a few hundred MB database is nbd.

I assumed that much.

I briefly considered spending $12k or something for the client-server programming course at ITT circa '98. Sooooo glad I decided to spend $78 at SF City College instead.

From this week's Hacker Newsletter: Are we in the Middle of a Programming Bubble?

Cliffs: "Is this real life? How can we get paid like elite professions like doctors and lawyers while putting in half the work? This can't go on forever, can it?"

It's definitely gonna contract one of these days when rich marks stop funding the 4th best of some new idea in hopes it gets bought by the 3rd string google or something. There is so much money right now that's just being fleeced from startup investors. I worked for one.

But we could have another social media type revolution that drags it out another decade+.

We probably are in a bubble but this seems to understate lawyer/doctor compensation. For example:

Starting total compensation for law school graduates at top law firms is over 200K+ and it goes much higher quickly. It's disingenuous to compare compensation at elite tech firms against compensation at law firms that aren't anywhere near top-tier. Lawyers do work longer hours at top firms, but after a few years, they can switch to cushy in-house roles that pay just as well or better than senior engineering positions at top tech firms while working standard 9-5.

After residency, doctors that work a lot of hours usually do so because the additional work pays so well and moonlighting is easy and well-accepted in medicine. Either way, doctors are compensated extremely well - this is from a survey done in 2015-2016 (so the numbers are likely ~15-20% higher now) and even include physicians working part-time (~16%), but the numbers are still very high across the board:

https://www.medscape.com/features/sl...verview#page=2

I don't think any average programmer survey ever results in numbers above $130K by comparison:

https://money.usnews.com/careers/bes...veloper/salary

Keep in mind that doctors are spread across the US, while well-paying tech jobs are concentrated in high cost regions. Also, it may take less effort to become a top software engineer than a doctor, but it's probably harder to keep the skills up-to-date as a software engineer.

Mostly I left programming for other reasons, but in the early 2000s I was thinking it was a little bubbly. Tech company fortunes went up and down I guess, but my friends in the industry from 1992 or so until now never had any trouble getting good paying jobs.

It's also important to realize that tech companies that are paying these high salaries are wildly profitable and their profitability comes primarily from extraordinary software engineering productivity. Top law firms aren't that profitable and it's pretty hard to tie their profitability back to the extraordinary work of their associates - to a large extent, they pay what they pay because they need to hire the best available to keep their reputation, not because they need talent on an absolute level.

High doctor pay is intimately tied to the fact that supply is artificially restricted and demand is price-inelastic due to the market structure and the unique inadequacy of the regulations in the US; even then, their pay appears divorced from their productivity:

https://www.nytimes.com/2019/01/24/b...ate-money.html
https://www.healthcarefinancenews.co...and-your-staff
https://www.forbes.com/sites/robertp...osing-millions

Meanwhile, given that the general plans at most top tech companies are to significantly grow their headcounts over the next few years, it's quite possible if not likely that the tech compensation overall will continue to grow at least in the medium term. It'd be interesting to see what happens in a major recession though.

Yeah it would. I've never seen a major tech bust so far, and that really includes the larger field of engineering as a whole in my lifetime, but allegedly there were some big problems in the 80s. You always hear about people going from like IBM to driving cabs, but it's hard to tell how wide spread that might have been.

It was alleged in the mid-90s that the reason that engineers were commanding such high salaries and hiring bonuses was that a lot of people permanently left engineering in the 80s and when hiring started back up, those people didn't come back and so there were shortages of experienced people. But again that's just the line I was handed in school so who knows.

The 2000 tech boom was definitely a bubble but most people seem to have walked away unscathed. But there may be vast reams of people who were barely qualified (or unqualified) for programming work who just went back to non-tech jobs. During that period we basically hired anyone who seemed smart, whether they had any programming experience or not, and basically ran bootcamps to train and indoctrinate them. It was a brutal time - we hired very readily but we also let people go after a few months if it wasn't working out. I think that's an effective way to hire, but essentially inhumane (and possible discriminatory, for example for foreign workers for whom losing a job could mean having to leave the country)

When I was 10 I sat next to a guy on a plane who used to be an aeronautical engineer until according to him the bottom dropped out of his industry. He was selling linoleum flooring or something.

I always think of that and never take it for granted that the gravy train is gonna be around forever. Although it's looking like it could outlast my working years.

Heh, I remember another guy on a plane who told me to look into XML as it was going to be the next big boom. I remember my non-tech, dotcom-investing uncle asking me too if I thought XML was going to be the next linux.

There really was a short window there where people thought you could somehow make tons of money off XML. Kind of like the 6 months or so where Facebook apps were going to be the next big thing.

From the article:

Quote:

Now, I LOVE these gigantic salaries. Even working at a non-FAANG company, my salary has been driven continually upward because of the pressure that Google and friends put on the programming market.

I just wonder if these numbers are sustainable.

Usually when I explain this phenomenon to people, they think it comes with a catch. They’ll say ‘Yeah but I bet they work like 80 hours a week’ or ‘Yeah, but I bet the work is hideous.’

From working there, that wasn’t my experience.

First on hours: I worked a lot during my time at Google, but I think I was the exception, rather than the rule. The only reason I worked that much was because I love working. Nobody was forcing me to do anything. I’d be at the office when the cleaning people came around, and usually I was one of the last 1 or 2 people there. Everybody else went home between 4 and 6 PM, depending on what time they arrived in the mornings.

FWIW I think it is pretty much accepted by posters on this forum that the concept of putting in your 8 hours and "clocking out" is ridiculous.

Quote:

Second, on the work: programming at Google is like enterprise programming anywhere. You do a lot of maintenance of legacy code, a lot of refactoring, and a lot of integrations. Every now and then you get to work on something really new and exciting. Assuming you like programming like I do, this is perfectly fine. I just love programming; what I’m programming on is mostly irrelevant.

To a large extent this is true but from ymmv regarding how dysfunctional the software development process is at a particular organization.

Performers at highly profitable companies get paid well. Seems fine to me.

Wasn't that toward the end of the period when anything computer related was getting huge coverage? XML was going to rule the world because it was better than HTML and HTML made the internet so presto-change-o we're getting a better internet!

Monday I'm getting an offer from a small MS Partner company doing support, they have a best in breed tool for a very large market and are planning on 50% growth this year so it there should be opportunities to advance. On first reading the job title and description I thought it could have been a jr product role which was why I applied. It's the first offer I've had since last January so I'm not going to worry too much that the pay is about half what I was making in Portland. It'll be enough to start filling in the hole I've been digging. Probably be a good excuse to cut back on 2+2 and their stellar 100 year old management style.