Hacker News Reader

Comment by stblack | Hacker News Reader

Maintaining and extending legacy code is, in many ways, orthogonal to its underlying language.

You can make a 2x2 matrix with "legacy code quality" on one axis, and "language percieved desireability" on the other axis.

You can make truckloads of money, earn much respect, with nearly zero negative ageism, in at least one of those quadrants.

"Building new" versus "maintaining existing" are different mindsets, and often a choice.

For older people just getting started, "maintaining existing" is often the wiser first decision. From there, the particular mix of language and tech stack count, and often the more obscure the better.

Find a good niche.

View on HN

Do you have any real world examples of this?

COBOL Devs probably earn more than React Devs.

Every time I've gone looking for COBOL jobs on job sites out of curiosity I always see them paying pretty poorly.

On Glassdoor for Cobol [0] the standard range is $54k to $179k (presumably in NYC metro where I am) with a reported maximum of $306k. The standard range for React [1] is $49k to $257k with a reported maximum of $550k.

I'm not saying Glassdoor is perfect but better to have these discussions based on some measured data than just common mantra that legacy is boring but pays better.

[0] https://www.glassdoor.com/Salaries/cobol-programmer-salary-S...

[1] https://www.glassdoor.com/Salaries/react-developer-salary-SR...

(seen from France)

Those are only public bids, and not everything goes through public channels.

Companies working with COBOL (or things like AS400) do so because they've been using it for some time already.

So they mostly already know the networks of professionals they will hire for the job, and that's often their own past employees who went freelance, or keep contracting after legal retirement kicks in. Professionals who also act as hiring agents for their replacements.

That's good to know. But for the purposes of people considering a new career wouldn't their position be basically the same as what you see publicly reported on Glassdoor? (assuming there's not some bias like people who use COBOL are less likely to also use Glassdoor and enter salaries there)

If you're starting a new career you wouldn't have a network.

If you're starting a new career, I'm not sure the info you will find on Glassdoor will speak much to you either (or, not as much as when you already have had one or two positions - because then you know how to interpret some signals companies broadcast through their ads, or their employees reports).

And looking after COBOL (or an old tech, without having been introduced to it before) as an entry job is a bit peculiar.

Better looking after the companies that still need it, understanding why, and then going through them for specific jobs, linked or not to COBOL.

I understand the appeal of getting into it for the technology or the money, but getting into it for the problems and how to solve them first will open many more paths (and as much if not more money).

How many of those React jobs include RSUs from privately owned companies? I would venture that approximately 100% of the COBOL jobs that's all cash compensation/salary.

Privately owned companies don't give RSUs normally they do stock options or revenue sharing but yes your point is good.

Startups and public companies both seem to pay better base salary for modern tech developers and also give you equity or profit sharing. Whereas the government or government contractor companies that often stick with old tech (it seems) you'll only get cash.

Banks may be an exception since many do give stock or good bonuses. But I don't often see bank positions looking for COBOL developers.

> Startups and public companies both seem to pay better base salary for modern tech developers and also give you equity or profit sharing.

I've worked for a couple public companies, none of which gave any stock or any profit sharing to the engineering staff. I've also worked for private companies that had some profit sharing but it was never discussed other than "there is profit sharing" even when it was a substantial percentage (one place I consistently got 6-8% of my salary as profit sharing).

More popular lately seems to be the pseudo-bonus-pseudo-profit-share thing where you have a set percentage of your salary as "bonus" but it requires the company overall hitting specific revenue and profit numbers, so it's essentially profit sharing with a cap.

I think my point is comp structures are all over the place, but "large public company giving base, bonus, and stock" is a small minority, while "private company giving base, and maybe a bonus sometimes" is the most common.

Why is that a bad thing to offer all cash? Equity from private companies is statistically worthless. I wouldn’t even count it as part of my compensation.

I don't think it's a bad thing (quite the opposite) but it's common to take high SV salaries that include hard-to-hit discretionary bonuses and hundreds of thousands in stock, and compare that TC number to the $175k offered by the private company that's also hiring. It's an apples to hammers comparison.

All of the public tech companies that I know about basically guarantee you cash + stock and a known vesting schedule over at least two years. Of course, we are living through the first era where BigTech stocks are tanking left and right.

I just left working at “normal companies” most of my career to work at BigTech two years ago at 46. My base pay was about the same. But my stock/prorated signing bonus brought my total compensation up by around 25%. I can (and do) sell my stock as soon as it vests.

I’m on the low end of BigTech compensation. I work remotely and did a slight career pivot from software engineering.

Yes, I know about the dot com bust. Those weren’t large profitable tech companies and they didn’t have the head count that companies have now

Super curious what your career pivot was.

Is there some engineering adjacent job that doesn’t come with a recruiter telling me to grind leetcode for 7-8 weeks for a chance to work on some weirdos space ship finance company? (Only kidding, mostly)

Cloud application development consulting in Professional Services - “application modernization”

> Every time I've gone looking for COBOL jobs on job sites out of curiosity I always see them paying pretty poorly

Sure, but the places known for paying well in the industry are also fairly overtly hostile to late-entrants and non-traditional career paths unless those people have already built a successful business and are getting acquihired.

Then more reason not to suggest people go into COBOL I guess (not saying you suggested they should, just responding about this thread overall).

Honestly, I really doubt it. This COBOL meme, is just that, a meme. Highest paid devs (>400k) generally work with brand new technology

COBOL devs have been at their same job for 10+ years. So take whatever the average dev wage was 10 years ago, and add 3% raises.

Zip Recruiter puts average salaries at around $88k/yr with the top 10% making over $110k. Payscale puts salaries a little lower than that.

The of the few job postings I see, none have salary information listed -- in my experience, that tends to mean the salaries are low or average. And if you're a COBOL engineer, jobs opportunities are slim, even in a region with a large number of older banks and insurance companies (read: prime COBOL industries). There's like one job each in areas like Omaha, NE; Alpharetta, GA; Montgomery County, AL; Chattanooga, TN; Milan, IL; Feather South, FL; etc. In other words, a bunch of places nobody has ever heard of.

The places that need to replace COBOL devs aging out of the work force with apps still running are among the places least likely to practice age discrimination to start with (and even if they were inclined to, there's not a pipeline of fresh out of college/bootcamp COBOL devs.)

(C#/Java have lots more opportunities at the same type of workplaces, so aren't bad choices but also lots more competition at all experience levels and from more traditional career paths; still, late-entrants with relative domain expertise can be quite competitive with pure programmers.)

I don't want to toot my own horn, but I wrote a blog post about being a niche programmer that you may or may not find useful: https://ano.ee/blog/the-niche-programmer

I am an extremely niche programmer myself so I'll check it, thanks!

Reply on news.ycombinator.com