whbboyd's recent activity
-
Comment on Career Advice for New Tech Workers in 2025 in ~comp
-
Comment on Career Advice for New Tech Workers in 2025 in ~comp
whbboyd (edited )LinkThis broadly jibes with my experience. I'll throw a few comments on, though: I actually disagree with this one. Having a portfolio is helpful to you; even halfway-decent interviewers will look and...This broadly jibes with my experience. I'll throw a few comments on, though:
Should I contribute to an open-source project to improve my resume?
- No. Companies that rely on open-source rarely contribute themselves, so why should you? Unless you want to be an unpaid idealist, skip it.
I actually disagree with this one. Having a portfolio is helpful to you; even halfway-decent interviewers will look and use it as a datapoint in the hiring decision, and it gives you something concrete you can talk about in an interview. (You'll want to talk about your previous work experience, if you have any, but you obviously can't share the code for that directly.)
That said, this can of course easily become a second (unpaid) job, which is rarely desirable. So I'd give the following tips:
- Ideally, your open source contributions should be of the "scratching your own itch" form. Contributing for the sole purpose of building a portfolio will probably burn you out, and probably be visible to employers, too.
- If you're just out of school, you probably have large programming projects for classes. Check with your professor first (the code you wrote is yours, but the problem statement and supporting material are not, and it may not really be possible to disentangle them), but putting these up in a publicly-visible place will be useful for at least a few years post-graduation.
- You don't have to contribute to large, established FOSS projects for this. A toy personal project absolutely has value in a portfolio. Spending a few hours a month on something fun or self-educational will give you a nontrivial chunk of code you can talk about and which demos your skills.
- Also, of course, you don't have to. Most programmers get hired with no meaningful public portfolio. If coding doesn't spark joy, definitely don't burn yourself out doing extra programming outside of work.
People can either be nice or good at their jobs, both are equally valuable.
This isn't a disjunction; you'll encounter plenty of people who are both nice and good at their jobs. These are, obviously, ideal coworkers, though less prevalent than everyone would hope.
The real condition here is that you can't be neither. People who are nasty and also bad at their job get fired. Talented assholes and friendly oafs might be on the chopping block for layoffs, but won't be proactively terminated.
Assume most of the decisions you are looking at were made by a person as smart or smarter than you and work from there has been a good mental framework when joining a team for me.
This is a good starting point, but a bad ending point. Everyone has cognitive pathologies that result in suboptimal code. Yes, even you! One of the most important things you can learn about your team is what these pathologies are, so you can judge decisions in light of them: was this historical decision made for carefully-considered (but inadequately documented) reasons, or because Senior Engineer Jay, who spearheaded the project, always makes that decision?
(An underappreciated downside to turnover is that when you don't know and can't talk to the person who committed a given headscratcher, there's no longer any way to unwind the decisionmaking.)
Seeing the profit numbers go up doesn't do anything for you.
There's an exception to this: if you like your team and job, your employer going under is obviously going to disrupt that. This is a long-term process at big corps, but can happen quickly—abruptly, even—at small companies or startups.
-
Comment on Journal that published faulty black plastic study removed from science index in ~science
-
Comment on Time for a new mouse? in ~comp
whbboyd Link ParentOh, I'm a terrible person to give advice about that, I'm afraid. I inherited a pro-quality soldering station from my electrical engineer grandfather, and learned the basics from my dad when I...Oh, I'm a terrible person to give advice about that, I'm afraid. I inherited a pro-quality soldering station from my electrical engineer grandfather, and learned the basics from my dad when I was—well, too young to be breathing flux fumes, anyway. =P
Ifixit has a general dissasembly guide for the G305, and switch replacement guides for other models. (To the best of my knowledge, all Logitech mice use electromechanically compatible microswitches, but that would be something to try to confirm before ordering anything.) Ifixit recommends this thing for the latter, which at a quick spec comparison is pretty comparable to the Pincecil V2, so the Pinecil will most probably work fine. If you're shopping around for irons, for a task like this one, more watts will always help—the big pins and pads make an excellent heatsink, and with too little power, you may struggle to get things hot enough to connect well.
-
Comment on Time for a new mouse? in ~comp
whbboyd Link ParentI can't speak for the G305 in particular, but Logitech mice in general are about as good as you could plausibly get on the repairability front. The generalized procedure is "unscrew the case,...I can't speak for the G305 in particular, but Logitech mice in general are about as good as you could plausibly get on the repairability front. The generalized procedure is "unscrew the case, desolder the malfunctioning component, solder in a replacement"—and that's going to be the procedure for something like a Ploopy, too. (Ploopy makes it easier to source replacement switches by having open parts lists, but to be honest, you probably don't want to replace Logitech switches 1:1—the OEM ones are, obviously, pretty bad, and when you're not trying to supply a whole manufacturing line, much better switches are very cheap.)
If you're not comfortable with soldering, there's basically no electronic repair above the level of swapping modular components you'll be able to do. And if you want to get comfortable with soldering, mouse switch replacements are one of the easiest things you can do: the pins are through-hole and enormous, and you can get very quick feedback on if your fix worked.
-
Comment on Lies I was told about collaborative editing, Part 1: Algorithms for offline editing in ~tech
whbboyd Link("Offline" meaning "reconciled later".) One very strong heuristic for the conclusion the author comes to via other means is: Software engineers have the best software tools. By definition, we can...("Offline" meaning "reconciled later".)
One very strong heuristic for the conclusion the author comes to via other means is:
- Software engineers have the best software tools. By definition, we can design and write them ourselves, precisely meeting our own use cases. The universe is not perfect, of course, so this is a rule of thumb and not, like, a law of physics or anything. But in my observation, it holds true in the vast, vast majority of situations.
- Because software engineering consists significantly of editing plain text, software engineers have the best tools for editing plain text.
- The tools software engineers use for collaborating on plain text editing—e.g., SCM like Git—never automatically resolve conflicts. Conflicts are flagged and sent up to the human user to resolve manually.
The heuristic suggests that if there were a better way to resolve conflicts than passing it up to a human, software engineers would be doing it; and we aren't.
-
Lies I was told about collaborative editing, Part 1: Algorithms for offline editing
14 votes -
Comment on How to judge relative dangers of chemicals for someone too busy (or lazy) to keep up with the science? in ~health
whbboyd LinkI really want a levelheaded analysis of the actual risks of microplastics, because all the rhetoric (and an awful lot of the research) about it triggers the same heuristics you mention for me....I really want a levelheaded analysis of the actual risks of microplastics, because all the rhetoric (and an awful lot of the research) about it triggers the same heuristics you mention for me. Like, microplastics cannot be "highly" toxic, because if they were, tons of people would be dropping dead from otherwise inexplicable poisoning (and would have been doing so for decades). They are at most subtly toxic.
The vast majority of the media coverage (at least that enters my bubble) is of the form "omg, microplastics discovered in <remote location or internal organ>!" This is a proxy endpoint from which you're supposed to infer that it's bad that they're there, even though the research doesn't establish that link. And in fact, like with aspartame, even separate findings of toxicity of the chemicals in question doesn't actually establish the harm from having them in <internal organ>. (Remember: treatments for petri dish cancer don't generalize.)
Anyway, my orthogonal rant aside, the best heuristic I have to offer is to think through the implications of the claim in question. If one of them is that a lot of poor, minority, or non-health-nut people would be dead or dramatically less healthy than they demonstrably are, the claim is either wrong or highly overstated.
(This was a helpful technique to reason about transmission rates early in the COVID pandemic. If you could multiply some numbers someone was panicking about and determine that a hundred million people should have died, those numbers were obviously wrong or not generalizable.)
-
Comment on How to judge relative dangers of chemicals for someone too busy (or lazy) to keep up with the science? in ~health
whbboyd Link ParentI mean, roughly zero cattle die a natural death. The vast majority are slaughtered shortly after adulthood, and most of the rest as soon as their non-meat utility (milk, breeding) starts to...0 deleterious effects
I mean, roughly zero cattle die a natural death. The vast majority are slaughtered shortly after adulthood, and most of the rest as soon as their non-meat utility (milk, breeding) starts to decline. An effect that shortens the average lifespan of cattle by fifty percent could easily go unnoticed.
-
Comment on What chemicals/substances do you keep at home? And what do you do with them? in ~life.home_improvement
whbboyd LinkWhat an interesting question! The following is from memory (and so probably incomplete), and doesn't include food. Kitchen: Distilled vinegar: for cleaning. It's great for removing mineralization...What an interesting question!
The following is from memory (and so probably incomplete), and doesn't include food.
Kitchen:
- Distilled vinegar: for cleaning. It's great for removing mineralization from kettles etc., and it's food safe because in a sense it actually is food.
- Liquid dishwasher detergent: just more convenient than powdered. Avoid single-use "pods", they tend to have a lot of packaging and include way, way too much soap.
- Liquid dish soap: for handwashing stuff that can't go in the dishwasher.
- Hand soap. My parents didn't keep hand soap by the kitchen sink, but dish soap is absolute murder on your skin, so I much prefer having it.
- Scouring powder (specifically Bon Ami): for serious, "I am not fucking around" cleaning. This stuff is a superpower I wish I'd learned about earlier in life.
- Ammonia window cleaner: I couldn't tell you why ammonia is the nasty chemical of choice for windows, but it is.
- Enzyme pet mess cleaner: for pet messes. Also works on potty-training-toddler accidents, ask me how I know.
- Food-grade mineral oil: for oiling wooden/bamboo cooking implements.
Bathroom:
- Assorted person cleaners: hand soap, body soaps, shampoos, conditioners.
- Porcelain cleaner, for porcelain.
- Isopropanol. Disinfectant, but it actually mostly gets used as a solvent. Stings like hell when used on cuts.
Basement:
- Laundry detergent. I use an ultra-concentrated liquid because it is vastly more convenient than other form factors.
- Liquid bleach: primarily for the washing machine's self-clean, though at this point my bleach is so old it's probably closer to brine than an actual cleaning product.
- Liquid drain cleaner: Dranō brand, lots of nasty stuff but mostly hydrogen peroxide. Don't use this, it's caustic, nasty to skin, and bad for your plumbing. My experience has been that blocked household drains are much, much easier to just mechanically snake with a cheap hand auger.
- Enzyme drain cleaner: Ideally I would religiously use a little bit of this on all my drains weekly, which would probably cut the amount of time I spend snaking them, but alas. =( If you need drain cleaner, this stuff is better all around than a caustic drain cleaner.
- Windshield wiper fluid: mostly methanol. I keep cold-weather formulations on hand due to the local climate.
- 3-in-1 oil: general lubrication. WD-40 is a shitty lubricant, 3-in-1 is better.
- Graphite lubricant: general lubrication.
- Cyanoacrylate glue: for when you just need two things (and also a bunch of your fingers) stuck to each other, dammit.
Shed:
- Motor oil. IIRC, SAE 10W-40? I had a car that burned oil. Now I don't, but I still have some spare oil.
- Gasoline. Not much, probably a half-gallon. 87 octane. Probably mostly sludge now. For a lawn mower I despised using, didn't use for years, and have now gotten rid of.
Notably lacking are any solvents stronger than isopropanol, or any acids stronger than distilled vinegar. None of my hobbies require them, and I haven't found a need for general household purposes like cleaning.
-
Comment on The affordable housing shortage is reshaping parts of rural America in ~finance
whbboyd Link ParentBoston is walkable. And, to address your point from another comment, at least when I lived there a decade ago, mass transit usage was very widespread and not really class-segregated. (Most of my...Boston is walkable. And, to address your point from another comment, at least when I lived there a decade ago, mass transit usage was very widespread and not really class-segregated. (Most of my coworkers used mass transit to commute, at least for the "last mile" segment downtown to the office if not the whole thing.) If you work in an office downtown, office events will be walkable or mass-transit-able away, and there will be a gaggle of attendees taking the subway you can ride along with.
Of course, Boston is in no way, shape, or form more affordable than NYC these days. And the experience of living there is very different (AFAICT, having only lived in one of the two cities) than NYC; it's much, much smaller and slower-paced. That could be good or bad, depending on your desires, but as an "NYC substitute" it is at best qualified.
-
Comment on The Texas OB-GYN exodus – Amid increasingly stringent abortion laws, doctors who provide maternal care have been fleeing the state in ~health
whbboyd Link ParentThese people obviously don't give one solitary shit about having a functional healthcare system.having a functional healthcare system
These people obviously don't give one solitary shit about having a functional healthcare system.
-
Comment on I am looking for 100% ad-free apps for older adults with dementia. Things like jigsaw puzzles, coloring and the like. Paid is fine. in ~life
whbboyd LinkNothing similar can exist on Apple until the EU finishes beating up on them, but on Android, you should basically never touch the Play Store and install apps from F-Droid instead ("solitaire", for...Nothing similar can exist on Apple until the EU finishes beating up on them, but on Android, you should basically never touch the Play Store and install apps from F-Droid instead ("solitaire", for example). It's entirely free software, which aligns the incentives completely differently. I don't think they have a hard rule against advertising, but I've never seen it.
-
Comment on Why is Google Gemini saying we should die? in ~tech
whbboyd Link ParentThis analysis is good, but I just want to point out: you are anthropomorphizing the shit out of the model, which in context, may not help the point you're trying to make. The model doesn't "want"...This analysis is good, but I just want to point out: you are anthropomorphizing the shit out of the model, which in context, may not help the point you're trying to make. The model doesn't "want" or "try" to do anything. It's a colossal matrix by which tokenized input vectors are multiplied. That matrix is a black-box locally-optimized (i.e. imperfect) solution to the problem of "given <training corpus, which in context broadly means 'the Internet'> and <prompt>, what tokens most probably continue <prompt>". That continuation is, of course, de facto heavily influenced by tone, and structural similarities (or lack thereof) to the training corpus, so the behaviors you're describing are real. But they're not "intentional" in any meaningful sense, just a side effect of leading a black-box model away from the data on which it was trained.
-
Comment on What is the best or recommended way to integrate my Windows 10 and Linux computers through the local network? in ~tech
whbboyd LinkI'll agree with the people saying: for your use case, it sounds like you want a NAS. This'll give you the ability to conveniently bulk transfer and store files, between arbitrary hosts on your...I'll agree with the people saying: for your use case, it sounds like you want a NAS. This'll give you the ability to conveniently bulk transfer and store files, between arbitrary hosts on your network if you use the NAS as an intermediary. You don't need specialized hardware (e.g. a Synology) for this; install network filesystem software on any computer with storage and a network connection, and tada! Network Attached Storage. =) In your case, it sounds to me like you could meet your needs by making your Plex box into a NAS.
The two main protocols for network filesystems are NFS and SMB/CIFS; either will work on both WIndows and Linux, but SMB will probably be easier overall if you've got WIndows systems in the mix. The SMB server software for Linux is called "Samba". For remote management of Linux/Unix systems (so you don't have to keep digging the computer out of its corner), everyone and their dog uses SSH; VNC or remote X (assuming it's using Xorg for its GUI, not Wayland) are possible, but would be very out-of-the-ordinary.
One thing to note about this approach is that you wouldn't have any redundancy in storage; one hard drive failure could potentially lose you a bunch of data (and with a convenient NAS, the temptation to centralize data will be very strong). If that's a concern for you, then it probably does make sense to invest in dedicated hardware for it. An integrated solution like a Synology will be very easy to set up, but you can probably save some money and get more control by throwing TrueNAS on a whitebox PC, or (if you feel really ambitious) building up from a scratch Linux install.
Another thing to note: don't limit your NAS to being just a NAS! No matter what form it takes, it's a server on your local network, and there's no reason it can't provide all sorts of services other than network filesystems. You already have Plex; depending on how you're doing your downloads, you might want to consider setting up Transmission's web UI. If you decide you want to explore home automation, drop Home Assistant on it. Et cetera!
-
Comment on Touchscreens are out, and tactile controls are back. Rachel Plotnick's "re-buttonization" expertise is in demand. in ~design
whbboyd Link ParentI still buy all my music on CD! Buuuuut… the first thing I do with it them rip them and put them on my phone, so the CD player in the car still doesn't get any use. =PI still buy all my music on CD!
Buuuuut… the first thing I do with it them rip them and put them on my phone, so the CD player in the car still doesn't get any use. =P
-
Comment on Advice Needed: Simple and Reliable notifications in ~comp
whbboyd LinkIs there a reason you can't use SES instead of Gmail to send your email alerts? For low volumes, the cost is a rounding error, and you can keep using pretty much all your exiting infrastructure;...Is there a reason you can't use SES instead of Gmail to send your email alerts? For low volumes, the cost is a rounding error, and you can keep using pretty much all your exiting infrastructure; SES supports SMTP for sends.
-
Comment on Best solution to extract PDF data? in ~comp
whbboyd LinkA former employer of mine built a tool for more-or-less this exact purpose: Textricator. Unfortunately, I was never involved in its use or development, so I can't help much with using it, but we...A former employer of mine built a tool for more-or-less this exact purpose: Textricator. Unfortunately, I was never involved in its use or development, so I can't help much with using it, but we were extracting structured data from court records which had been rendered as PDFs, which sounds reasonably close to what you're doing.
-
Comment on ‘I grew up with it’: readers on the enduring appeal of Microsoft Excel in ~tech
whbboyd (edited )Link ParentNo structure. This doesn't matter (and is potentially kind of an asset) when human beings use it as computerized graph paper, but it is an enormous fucking pain in the ass when they then decide...- No structure.
This doesn't matter (and is potentially kind of an asset) when human beings use it as computerized graph paper, but it is an enormous fucking pain in the ass when they then decide they need their
abominationdata liberated and hand it off to me to write software to process, which always happens eventually for any sufficiently long-lived spreadsheet.Some (almost certainly not exhaustive, I've blocked plenty of trauma from my memories) examples I've personally run into:
- Excel actually allocates cells, and it does so lazily, and there's a failure mode where it will allocate the bottom-right cell in a sheet for no reason. If you're iterating rows (which is basically the only reasonable way to try to get structured data out), you'll get your data rows, then a million empty rows, then the very last row with one empty cell in it (in a column you're certainly not looking at). This will either break your processing or just waste a ton of time, depending on how much effort you put into handling ill-structured data.
- Also, oddly enough, a hair over a million rows is not enough for lots of datasets. (And if, god help you, someone is slinging around the Office 2003 format, that has a limit of 64Ki, which is not enough for any dataset worth wasting a software engineer's time on.) It is oddly easy to silently truncate when you hit the row limit, as very infamously happened to England's public health service with COVID data.
- If there are dates in a spreadsheet, they are wrong. You should basically just count on it. Excel's date recognition and mangling is tailor-made to corrupt dates.
- Because Excel goes to great lengths to infer cell types, you can't round-trip arbitrary strings through it. It will decide some strings are a different type, convert them, then format them differently on the way out. (This is far and away most prevalent with dates, but definitely affects other types, too.)
- Of course, the data in any individual cell is totally free-form. You have a sheet that looks like a big table, with one column that looks like it's got numbers in it, but if your software isn't prepared for one of those cells to contain an embedded rickroll, it's going to crash. Excel has this nice feature where you can set a limited list of possible values, but aside from being per-cell (and thus almost certainly inconsistent), it's part of the cell's format, purely advisory, and you should probably count on some cells having data that doesn't conform.
-
Comment on Morrowind doesn't have any rivers in ~games
whbboyd (edited )Link ParentOh, no, I get it. My reply is partly in kind; but I do think the video suffers for being too straightfaced. The premise is ridiculous; it needs at least some nod to that ridiculousness, or it kind...Oh, no, I get it. My reply is partly in kind; but I do think the video suffers for being too straightfaced. The premise is ridiculous; it needs at least some nod to that ridiculousness, or it kind of reads like the author is just… kind of a jerk about the limitations of 2002 videogames.
(I watched the author's Skyrim waterways video, and IMO it works a lot better, because it breaks the "fourth wall" of game design in its analysis. Why does Skyrim have a river which reverses direction at one point? Because it's a videogame! There's no actual physics acting on the water, and the designers just… needed it to be going the other way.)
Also, if you really get into the weeds, assuming the game is the platonic ideal of Morrowind is self-contradictory, since the in-game lore plays with the concepts of the "player character" and the game as window into the world. But that's some real lore dweeb shit; I'm not judging anyone for staying far afield of that rabbit hole.
While this is possible, don't assume it. Every application is a long shot; hiring rates for junior and mid-level positions are absymal right now (I've heard from colleagues of listings which got literally thousands of applicants, a sub-0.1% overall acceptance rate, although obviously almost all of those applicants never interviewed). If you assume a generous 10%—i.e., ten applicants to every opening—you would, given random selection of applicants (which there obviously isn't, but it gives an idea of a baseline), need to apply to seven jobs to beat even odds of acceptance, and 22 for 90%. To reiterate a point from the article, failure is the normal outcome of a job application. Try not to let it get to you, and try to be as objective as you can in your evaluation of your application and interview performance. (Also, if you do interview and are later declined, ask them why and if there's anything you should have done differently when applying.)