|
Y'all are misunderstanding. You ultimately NEED to measure individual performance somehow, because it's the individual that gets the money. I don't think we disagree here. But you can't measure individual performance against a team's KPI, or really any one set of numbers-- which is what some posters in this thread are suggesting, or what some bad managers who create bad reward systems try to do. The point I'm making is (1) the manager is always evaluating individual performance against individual expectations, and (2) the manager needs to spend time figuring out individual expectations (and the signals for those expectations-- because expectations are abstract qualities). As an aside, it's the job of the manager to at least have a sense of their reports' contributions to a group project, whether or not the individual manager is the team manager. Pie Colony fucked around with this message at 00:30 on May 4, 2019 |
# ? May 4, 2019 00:28 |
|
|
# ? May 30, 2024 23:35 |
|
https://twitter.com/lhochstein/status/1123755527312433152 Mine is "self-inflicted DDOS", but that's not a phrase that comes up too often.
|
# ? May 4, 2019 00:54 |
|
ultrafilter posted:https://twitter.com/lhochstein/status/1123755527312433152 "CAP Theorem" or "high availability"
|
# ? May 4, 2019 01:08 |
|
ultrafilter posted:https://twitter.com/lhochstein/status/1123755527312433152 Is "technical debt" too easy? I feel like we all must have some horror stories about really, really bad code we've had to work with.
|
# ? May 4, 2019 01:44 |
|
Forget testing, fix it now!
|
# ? May 4, 2019 01:47 |
|
TooMuchAbstraction posted:Is "technical debt" too easy? I feel like we all must have some horror stories about really, really bad code we've had to work with. "production outage" "marketing/sales promises" "stop gap measure" "proof of concept in production"
|
# ? May 4, 2019 01:49 |
|
Zero downtime deployments
|
# ? May 4, 2019 01:55 |
|
New Yorp New Yorp posted:"production outage" This one reminds me of an amusing story from when I was working at Amazon over a decade ago. They had plenty of monitoring on all kinds of stats, but the big one was order rate, i.e. the rate at which people were making purchases on the site. If the order rate dropped out of the expected bounds, just about everyone got paged. Once we got paged because both the UK and German sites had experienced massive order drops. Nobody could figure out what was going wrong. About an hour into the order drop, suddenly things returned to normal (again, with no reason as far as we could tell from looking at our metrics)...then they dropped again a few minutes later, and returned to normal again an hour-ish after that. It turned out that the UK and German football teams were playing a match in the World Cup, and practically everyone in both countries was tuning in to watch, instead of shopping. Except during halftime, of course.
|
# ? May 4, 2019 01:56 |
|
TooMuchAbstraction posted:It turned out that the UK and German football teams were playing a match in the World Cup, and practically everyone in both countries was tuning in to watch, instead of shopping. Except during halftime, of course.
|
# ? May 4, 2019 03:06 |
|
Pie Colony posted:Y'all are misunderstanding. You ultimately NEED to measure individual performance somehow, because it's the individual that gets the money. I don't think we disagree here. But you can't measure individual performance against a team's KPI, or really any one set of numbers-- which is what some posters in this thread are suggesting, or what some bad managers who create bad reward systems try to do.
|
# ? May 4, 2019 03:44 |
|
ultrafilter posted:https://twitter.com/lhochstein/status/1123755527312433152 "benign race"
|
# ? May 4, 2019 03:46 |
|
ultrafilter posted:https://twitter.com/lhochstein/status/1123755527312433152 my most hated feature request in the planning stages: "simple ad-hoc reporting" fuuuuuck you, here's a CSV
|
# ? May 4, 2019 13:49 |
|
JehovahsWetness posted:my most hated feature request in the planning stages: CSV is very good because when they say, "couldn't it do X easy ad hoc analytics for me at my request?" you can say, "Yes! Right after you load it into Excel or any of the other familiar analysis programs that it is completely fruitless for us to partially reimplement!"
|
# ? May 4, 2019 13:55 |
|
We recently built a fancy dashboard that the product manager thought would be easy to implement. Then we sat him down and told him that, no, none of this is easy, and would probably take several times what his initial guess was. Deciphering his specs (which were, to be clear, extremely lovely and underspecified -- basically a mockup with a few comments) took two weeks and a lot of teeth pulling, for a project he assumed would take a month total. I don't expect product managers to be technical or to even understand (or care, really) why we make particular engineering decisions, but I do expect them to understand that things are never as easy as making a mockup and telling us to build that and assuming it's easy. Especially when they've been here long enough to know that nothing is easy with a lovely legacy code base.
|
# ? May 4, 2019 15:23 |
|
The best product managers I've had have had an engineering background to begin with and could much more accurately: 1. Estimate the scope of work involved without constantly holding meetings 2. Call out other engineers on BS over-estimation A small POC doing stuff that takes like a day for any engineer on the team to do shouldn't get porkbarreled into a ticket where someone fights with Jenkins job refactoring for 2 days. With that said, product managers that move in from engineering seem to have a fair bit of trouble prioritizing the tickets that really make everyone more effective, but this to me is more indicative of being younger / less experienced in software product development rather than something endemic to starting out in engineering.
|
# ? May 4, 2019 17:03 |
|
"Can we outsource this?" [understood to be said in a manner so eager that it makes you want to smack the guy who said it]
|
# ? May 6, 2019 14:41 |
|
ultrafilter posted:https://twitter.com/lhochstein/status/1123755527312433152 "We'll fix it after release"
|
# ? May 6, 2019 19:08 |
|
Is there a good resource for learning about "software testing patterns" and "software testing best practices"? I know about software design patterns but I've never come across testing patterns before.
|
# ? May 6, 2019 19:43 |
|
Uhh Nope posted:Is there a good resource for learning about "software testing patterns" and "software testing best practices"? "Software testing" is a huge topic. What kind of testing are you talking about? Unit testing? Integration testing? End-to-end testing? Manual testing? Load/performance testing? Fuzz testing? Security/penetration testing?
|
# ? May 6, 2019 19:51 |
|
New Yorp New Yorp posted:"Software testing" is a huge topic. More then a few of these can be broken down into further categories. I would like to add: test specification techniques, static testing, monkey testing, system testing, contract testing, mutation testing and stress testing. Some are intentionally overlapping with some of yours. Personally I think the core of what is considered testing in the dev world is here: https://martinfowler.com/articles/practical-test-pyramid.html but Uhh Nope should just explain what it is that (s)he is trying to achieve and then we can first tell her/him it is wrong and then tell another solution that will not work in their specific scenario.
|
# ? May 6, 2019 19:58 |
|
Keetron posted:More then a few of these can be broken down into further categories. I would like to add: test specification techniques, static testing, monkey testing, system testing, contract testing, mutation testing and stress testing. Some are intentionally overlapping with some of yours. I will definitely read the article you linked. In the meantime I think what I'm looking for is performance testing, monkey testing, and integration testing, those three topics.
|
# ? May 6, 2019 20:05 |
|
I would really dig seeing a good article about varieties of software testing written more from a business perspective, i.e., something that matches particular types of use contexts and customer relationships to kinds of testing.
|
# ? May 6, 2019 21:41 |
|
Keetron posted:Personally I think the core of what is considered testing in the dev world is here: https://martinfowler.com/articles/practical-test-pyramid.html Wow that article was super helpful, covers lots of ground. That said I guess is there anything else I should know? It didn't really get into "patterns" at least not how I recognize patterns to be, I guess CDC tests are kind of a pattern but it's not exactly what I expected like "Flyweight Pattern" or "Strategy Pattern"
|
# ? May 7, 2019 18:39 |
|
You can look into BDD: https://en.wikipedia.org/wiki/Behavior-driven_development that is supported by https://cucumber.io/ This is a technique that is aimed at moving forward with the desired business outcome as a starting point. On test design patterns, for unit tests there is this book that is on my shelf but unread so far: http://xunitpatterns.com/ I working in the testing / QA domain from 2004 to 2016 in a variety of roles and am currently at the point that most of the time, a dedicated tester is not hired to make sure the software is delivered as intended (that is a very deliberate choice of words) but is hired to cover management asses for not taking care of the development process properly. This management of development is rather hard so instead of doing hard work, the decision is made to hire testers and then blame them for any screw ups in prod. I am not bitter, why do you ask? prisoner of waffles posted:I would really dig seeing a good article about varieties of software testing written more from a business perspective, i.e., something that matches particular types of use contexts and customer relationships to kinds of testing. On my current project I have been asking for user involvement to start about 6 weeks before we would deliver to a single user. We are planning to move to prod in two weeks and have a first meeting with a user that is not the PO in two days. Not sure where I am going with this, other than saying that testing will not solve any problem that a lack of collaboration caused.
|
# ? May 7, 2019 20:42 |
|
Keetron posted:I working in the testing / QA domain from 2004 to 2016 in a variety of roles and am currently at the point that most of the time, a dedicated tester is not hired to make sure the software is delivered as intended (that is a very deliberate choice of words) but is hired to cover management asses for not taking care of the development process properly. This management of development is rather hard so instead of doing hard work, the decision is made to hire testers and then blame them for any screw ups in prod. I am not bitter, why do you ask? Keetron posted:Not sure where I am going with this, other than saying that testing will not solve any problem that a lack of collaboration caused. strong, strong agree.
|
# ? May 7, 2019 21:34 |
|
I am just gonna say that xunit frameworks are terrible, but I guess that comes with the language that originated them.
|
# ? May 8, 2019 08:21 |
|
How in the world do I figure out how much I should be making in my local market? I live in St Louis and make 88k wth 5 years experience. I just got a LinkedIn message from a recruiter about a local position that would pay 120k with a 10% bonus, which has me thinking maybe I'm underpaid.
|
# ? May 8, 2019 20:42 |
|
Doghouse posted:How in the world do I figure out how much I should be making in my local market? I live in St Louis and make 88k wth 5 years experience. I just got a LinkedIn message from a recruiter about a local position that would pay 120k with a 10% bonus, which has me thinking maybe I'm underpaid. Go to interviews, don't tell them how much you're earning right now, look at what offers you get. By doing this you're also getting some free interview practice and maybe even a chance to learn something interesting about how things are done at other companies. I think it's a good idea to do interviews all the time, even if you have no interest in changing jobs.
|
# ? May 8, 2019 20:49 |
|
Doghouse posted:How in the world do I figure out how much I should be making in my local market? I live in St Louis and make 88k wth 5 years experience. I just got a LinkedIn message from a recruiter about a local position that would pay 120k with a 10% bonus, which has me thinking maybe I'm underpaid. You should be making as much as you can get someone to pay you, local market be damned. If you can get an offer from a place that's willing to pay you 30k more than you're making now, then you aren't necessarily being underpaid. But you should still strongly consider it. The best way to see what's "normal" for your area is to look at companies on Glassdoor and other salary-aggregation sites, and also look at local job postings similar to what you make now with salary ranges included. Assume the higher end of the salary range.
|
# ? May 8, 2019 21:00 |
|
sunaurus posted:Go to interviews, don't tell them how much you're earning right now, look at what offers you get. Do you use vacation time for all these interviews?
|
# ? May 8, 2019 21:46 |
|
Pretty sure anonymized h1b salary data is freely available and indexed on various sites on the web. Look for a midsize local company.
|
# ? May 8, 2019 21:48 |
|
Asking friends and people on here is good. Also, when you're working for a company, ask your boss about the salary bands, or look through the company's publicly shared folders.
|
# ? May 8, 2019 21:57 |
|
dantheman650 posted:Do you use vacation time for all these interviews? I just checked, I went to 5 interviews in 2018, 4 of those were an hour long and I did them during lunch, only the fifth required me to shuffle my working hours around a bit, and I've definitely never used vacation time for interviews. Are you implying that doing interviews is some big thing that requires a lot of invested time? I've heard about FAANG etc having insanely long and involved interview processes, but I've never experienced anything like that myself. Maybe that means I only get invited to interviews at crappy places.
|
# ? May 8, 2019 22:01 |
|
Every on-site interview I’ve ever done has been minimum 3 hours without travel time. What companies are you interviewing at that only take an hour? Do they only have you talk with one person? That seems far outside the ordinary from a few years of reading this thread and anecdotal experience. Most on-site interviews are 3-4 45 minute coding sessions with 1-2 shorter behavioral sessions, plus a food break.
|
# ? May 8, 2019 22:16 |
|
Yeah I think the shortest interview I’ve had was a half-day.
|
# ? May 8, 2019 22:43 |
|
The only time I've had to write actual code for an interview was when I was applying for my first job as a junior dev, and even that was a take-home thing. After that, I've interviewed at startups, small and medium companies (the place I work at right now has 5000 employees in 17 countries, I guess that's medium?), and none of them have ever asked me to write code for interviews. The most common interview scenario for me has been something like ~15 minutes of me elaborating on my CV and answering questions about myself and my experience, ~10 minutes of somebody telling me about the company, and then ~30 minutes of free-form discussion about the team I would be joining, where basically both myself and the interviewers are trying to figure out what the other side thinks about culture/technologies/processes etc that are relevant to the team. Usually I'm interviewed by 2-3 people at a time but I've also had interviews where the whole team was present. A few times, I've been invited back for more 1-hour interviews before getting an offer, but mostly I seem to get an offer after the first interview (usually along with an invitation to meet the rest of the team). Is my experience unique here then? I can't imagine investing a whole day just on interviewing somewhere unless I already know I really want to work there.
|
# ? May 8, 2019 22:58 |
|
The longest interview I've had was 3 hours. I goy the job and quit after 4 months because it was horrible
|
# ? May 8, 2019 23:04 |
|
I just did a 3 hour interview yesterday. Not writing code, but a fair bit of whiteboarding stuff.
|
# ? May 8, 2019 23:20 |
|
sunaurus posted:The only time I've had to write actual code for an interview was when I was applying for my first job as a junior dev, and even that was a take-home thing. After that, I've interviewed at startups, small and medium companies (the place I work at right now has 5000 employees in 17 countries, I guess that's medium?), and none of them have ever asked me to write code for interviews. Where are you? That's definitely unusual for large parts of the US.
|
# ? May 9, 2019 00:28 |
|
|
# ? May 30, 2024 23:35 |
|
leper khan posted:Pretty sure anonymized h1b salary data is freely available and indexed on various sites on the web. Look for a midsize local company. Whoa. I never knew this, that's good to know. Thanks for the feedback, all.
|
# ? May 9, 2019 02:14 |