Theory: The software architect is a super-seasoned dev-lead who has learned by past fiasco how to avoid a large range of mistakes. The architect is also aware of all initiatives going on across mega-corp, and thus can drive alignment and efficiencies.

Reality IME - Scenario #1: The architect has enforcement power but no involvement in implementation. The architect crams a bunch of fancy-sounding unnecessary crap down the implementation team's throats. Eventually the implementation team removes the unnecessary crap, but not w/o a lot of pain, and not until after the architect has claimed credit and moved on.

Scenario #2: The architect has no enforcement power and no involvement in implementation. This was me for 2 years. The implementation politely team sits through a 2 hour handoff, then immediately throws away everything they don't like - which is usually everything.

Actual outcome from scenario #2 is that a project that was originally planned to serve the whole spectrum of dev teams (like a "universal cache") ends up serving only the team who got their first, on their servers, with their custom serialization code - inaccessible to anyone else - include node entirely, which was supposed to have been a requirement from the beginning. Now if I was more seasoned as to how mega-corp worked - I might have been able to twist the right screws and alter this outcome. But my only instructions were to sit on meetings and offer insight.

I'm sure it works better in other places - but if the "architect" (or whatever the role is called) isn't on board through the first production release, and several more until maturity, then the model completely fails imo and the architect is actually a huge liability.

This is pretty much exactly what our principal architects do FYI. They can encourage some direction or need, help define MVP on stuff, and make design decisions when needed (what db, what framework, blah blah). Im building an SSO integration currently and mine recommended checking out OpenId Connect instead of SAML, which is 100x easier and way nicer to work with. Seems fairly valuable.

Tim,

This isn't directly addressing your question, but I work for an IAM consultancy that does a lot of Sailpoint work. Remote/WFH when you're not doing client travel. Shoot me a PM if you ever have any interest in that, or just want to talk shop re: IAM.



Sent from my SM-G950U using Tapatalk

@KatoKrazy - congrats on your offer. I get what you are stating about that company wasting your time. Knowing the kind of work you are involved with I think you'll find companies are facing extremely severe challenges in meeting their project milestones and controlling costs. I don't think the architect role is relevant anymore in the firmware development field so I am thinking maybe the one offer you are getting is the best anyway.

Lol Yahoo always and forever

Dang it. I set up a code analysis tool (sonarqube) on one of our repos, to see what the analysis was like. Most of it isn't too bad or is good advice, but man, some of it is draconian. Like, your functions can't take more than 7 params. There are looooots of functions of mine that do, because they are essentially sql queries and every one of them takes a bunch of named params that control how they act, like
sort
page
limit
include_count

and usually they take several variables that control what's in the where clause.

So now I guess I have to manually mark them all as false positives or something.

I mean I guess I get the idea but having a bunch of standard optional parameters doesn't seem like a problem to me.

It's mad because one of my source files has more than 1000 lines

Instead of 50 flags as params cant you just make them one integer that gets OR’d with whatever you need?

So first of all, those aren't all flags. It would be like
do_some_query(user_id=2, page=1, limit=20, sort='username', include_count=True)
and so forth. do_some_query will probably take a half dozen argument that can be used to limit the results, like say user_id, username, email, etc. All query functions take at least the page/limit/sort/include_count params so it's really easy for functions to have more than 7 args.

Also I would say that if I was really just taking a bunch of flags, using distinct params is still better because the function prototype is self-documenting.

both tells you what parameters are available and what the defaults are. If they instead took flags, you'd have to look deeper into the code or somewhere else that the constants are defined. That kind of thing makes a lot more sense in languages that don't support named parameters, like C.

Ah, my bad.

I personally hate funcs with a lot of params. If i found myself doing it a lot I’d be really likely to try to make some type of struct/object/class/whatever floats your boat and pass that in.

Probably doesn’t make sense for your use case though

Mucho named params >>>> much positional params. But yeah I tend to do an object to.

The Sonarqube battles at my company between DevOps and devs were hilarious. At some point DevOps realized the devs had hidden like 70% of the code base from Sonarqube. I looked at the JS stuff it found and the issues seemed pretty legit. But the devs had a reason for each thing - mostly being massively behind deadline.

Thanks.

It's not over yet though. The bureaucrats in HR got their grubby hands involved and made the offer for 1 level lower than we had discussed. All those idiots seem to care about is years of experience, rather than what you have accomplished in those years and what your skills are.

Consequently the offer was about 15-20k less than I was expecting. I'm certainly not going to be moving my family across the country for a lateral move.

Thankfully I have the leverage here as the manager and team is familiar with my work, the work that they are asking me to do is definitely above the pay grade the offer was for, and I don't have to move as I already have a good job.

I talked to the manager and the recruiter and told them to work it out and get back to me if they can give me a fair offer, so we'll see what happens.

Again, I program in R. I like it. But I'm not a comp sci guy, I'm not a real programmer or whatever. From my perspective, I totally agree. I like R quite a bit, but there's this "..." argument in all these functions that is basically the Wild West in that all sorts of random, not-explicitly-named arguments may apply.

I have another not-a-real-programmer question. So R is object oriented. Okay. Then, there's functional. I have read about the differences, but I just can't grok what functional programming is. I've looked at which languages are considered functional and they all seem like lightly used languages compared to something like JS. Wtf IS functional programming?

Well I understand your frustration and far be it from me to defend HR types but they do have to compare your credentials with the credentials of people that they employ already.

Might want to consider MSFT if you are looking for a good pay day. They actually employ lots of firmware developers. Beautiful campus, nice work environment and they don't seem to be as encumbered by HR types.

https://careers.microsoft.com/us/en/...mware-Engineer

I can definitely see where you are coming from.

I'm not looking to be paid a premium, I'm just looking to be compensated fairly for the work that I will be performing. The work would actually probably fit a couple levels above what I am even seeking.

For example, the team I am on right now is an Engineering Fellow, a couple of Senior Technologists, and a few Technologists, and then lowly me. I am on the team at the request of the Fellow and one of the Senior Technologists. I am preforming way, way above my pay level. I really don't want to be in that situation again.

The best time to get what you want is when transitioning jobs, so I'm definitely going to go after it. I have quite a bit of leverage here, I certainly don't have to pick up my family and move halfway across the country. If we can't come to an agreement, no skin off my back, right?

so i start full time this week officially and I think it's going well, I'm starting to really steer the project direction and dictate/prioritize what we're going to work on (with my boss's blessing and everyone's agreement). Have identified a lot of issues that are related to another that werent really noticed by the team before, so right now I'm kind of functioning as a "librarian" for each project I'm working on and pointing out "oh, this depends on that, you'll want to do x first before we work on that" etc.

Right now i have 2 projects but I share the same team for both of them. One is new development from the ground up basically but I don't really have any deadlines to shoot for. Everything's kind of lackadaisical.

What I'm really struggling with right now is how to fill my day. Answering emails, scheduling meetings, and tinkering around with crap on github only takes up a few hours of my day tops - the rest of it I just go over the projects and my notes and make sure I'm up to date on everything but honestly that doesn't take long either.

Now that I'm there 8 hours a day I really need something to fill my time other than just browsing code to better understand the architecture. Today I implemented a simple malloc implementation in C so I could better hone my memory/pointer skills. Learned some keyboard shortcuts for my mac because the girl on my team seemed really annoyed with how slow I was with my presentation today and commented how I need to learn some shortcuts. But most of the time I seriously don't know wtf with myself. How do you guys fill your day??

I made a rough outline of a project roadmap and think I should run it by my boss but that doesn't take that long either. I also started estimating costs (in man hours) of stuff I thought was wasting a lot of time and effort. Maybe I'll work on a presentation for that or something.

My boss is gonna be gone for all of july so I think it's going to fall to me to make a lot of decisions in that time, maybe there will be more to do. I feel ready but I don't have a solid enough grasp of the finer details of the project yet, so I guess I could work on that more. But really i feel this job isn't that hard and not a lot is expected of me so I'm not sure what to do from here.

uh we write code mostly

A lot of full time jobs simply don't have 40 hours of work to be done each week. You choose between getting creative/proactive or slacking.

And prepare for the occasional week where **** hits the fan and you have 80 hours of work.

Yeah that too

His job is basically making sure the devs have enough work lined up that there is always something they can be working on.

And to make sure this isn't the new normal.

jmakin, what do the stories in your queue look like? Do you have a definition of what a story needs to have to be considered complete? Do they have enough information for the dev? Are they in the right format? Ours were written in a "Given x Then y" format, basically BDD statements that where tests to know if the story was complete. Getting your backlogs into shape should keep you fairly busy.

If you still have time try using your team's velocity and the backlog to validate your delivery schedule.

You could also be working on ideas to present to the team for defining what "done" means. Is it code merged back to master, is it code tested in dev ready to be merged? Researching this can find work that is hidden because it is assumed to be done by someone else.

I need some advice.

I've been interviewing with a company that's interested in hiring a senior engineer. I'm relatively happy at my current job but this company seems pretty interesting and might be a better fit for me, but that might also just be grass-is-greener syndrome.

I just finished the main interview and they told me I would be contacted by their hiring manager so I assume he'll negotiate for a salary. When he initially contacted me he mentioned they are currently paying around 8.5k/mo for senior engineers. That's slightly lower than my current salary, and honestly I think its below market. I have some research that says the median for someone with my experience is ~9.6k. That is total compensation however, and this company does offer an 8% performance related bonus on top, but I don't know how often they pay it out so I can't really estimate how much compensation that would be.

So, I'm thinking of telling them that I want 11k base. My reasoning is that I need an incentive to move from a position that I'm satisfied with, and I have a history of settling for lowball offers and I want to stop underselling myself. Do you think that is too high of a counter? I'd also like to keep the option of working there in the future open, in case there are changes in my current work.

I also have a feeling that they'll counter by mentioning the bonus, how do I deal with that? Should I ask for an estimate of how often that bonus gets paid out and calculate the EV of that as compensation towards my target?

I’m basically using github projects to manage everything which is just the worst but my hands are sorta tied there. My stories are in the form of issues that I usually break into smaller pieces if possible.

They’re a little hard to define because our projects are not even pre-alpha yet, we have some licenses but they are very dumbed down versions of what we spend most of our time developing. They’re enormously complicated.

For an “issue” or story to be complete one of the first things I required was that it’s merged to master. One of our devs had a habit of making a new branch for every fix he made and then forgetting about it, so he had the same ****ing fix spread across a dozen different branches. I nipped that **** in the bud early.

For a story to be done it needs to be tested, pulled, reviewed, then discussed at the sprint recap. For simpler ones it’s not as stringent.

I review and sort and categorize the stuff in the backlog for about an hour a day. I probably need to start adding stuff to it. One of the things I need to work on right now is a project roadmap.

I made an hour by hour schedule for myself today and that’s working out a lot better.

Who pays monthly?