4 ms·

As one of the 99%. What are some of these problems exactly, that me and my extremely large toolkit of cool things I can curl from Github can't solve? Because

by urthor 5y ago

As one of the 99%. What are some of these problems exactly, that me and my extremely large toolkit of cool things I can curl from Github can't solve? Because I'm not a genius. But I'm real good at stapling together some useful crap.

oblio 5y ago
You'll probably get a long list of stuff that includes things such as: OSes, compilers, game engines, distributed process frameworks, etc. Which as you'll note, are infrastructure things. Things which need 1/100th the manpower (at most!) to maintain. While everyone else is busy churning user-facing apps. You don't need to do that stuff, you career can be extremely successful without ever going near those things, especially if you are not interested in them.
- urthor 5y ago
  That's the thing, I think a lot of those are, if not solved, problems where there's a lot of really well established good solutions that exist. I could, I guess, tackle them. But I've got an appreciation that what's already there is already pretty great. I suppose that's maybe the answer to my question. You don't need to be 1% to solve a problem as well as the solutions that already exist. You need to be the 1% to improve on the current state of the art. Which in a commercial environment, as you can imagine, is exceptionally rare. They are more about finding unsolved problems than iterating on problems already solved by FOSS.
  - oblio 5y ago
    For various reasons (usually control, but frequently spun as "needing to be cutting edge"), some groups or companies do reinvent those wheels and in some lucky cases we get progress. For example Google has recreated several of those things and especially in the data storage field they've probably moved things forward. But in general I agree with you, the crushing majority of developers out there won't work on those things and won't ever even need to. It's also a source of not always deserved elitism from the people that work on those things, and I definitely dislike that (there's enough elitism in this field, as-is).
    - urthor 5y ago
      Elitism is driven by ego and self actualization though. I think there's a lot of people who are highly, highly motivated to work on tough problems with competent co-workers. Those emotions are less driven by the domain itself, and more about the fact that those people have succeeded in finding meaningful work with co-workers of appreciable worth. The tricky thing I find is finding that. Far too many regular projects are focused on mobilizing the dumbest blocks of wood towards a goal. Without actually investing in the human beings doing the work. I think once we regain a culture of actually investing in people who aren't in the top 10% in their field, things can change. Currently though we do not have that culture.
themacguffinman 5y ago
Building a high availability system? The downtime (especially scheduled downtime) of many of these "99%" services is shocking compared to the cutting edge consumer services we're used to. The "hug of death" is an affectionate meme that basically describes the remarkably common scenario where a "99%" website goes down at the precise moment at where it's peaking in popularity and profitability. Or how about high performance? Google famously knows that every millisecond of delay strongly impacts user retention and bounce rate, yet you wouldn't know it browsing many "99%" services. Achieving consistently low end-to-end latency is really hard regardless of scale, it's so easy for performance issues to creep in from every corner of your stack and accumulate. These are hard problems and IMO they are rarely solved well by just stapling together some useful crap.
- falcolas 5y ago
  Those are usually business decisions. Pay an extra $$$ for an outage scenario that occurs once a month, and doesn't materially impact our users? Hell no. Spend XXX developer hours and/or $$$ on hardware to make the site 50% faster, instead of adding new features our users pay for? Hell no. And let's be honest, for touting speed, Google's awfully slow and bloated these days.
  - themacguffinman 5y ago
    A hug of death when your product has peak user interest...doesn't materially impact users? Your users don't care about multi-hour or even multi-day downtimes in "99%" industries like...healthcare and banking? What a perfect demonstration of "Those scenarios are rare if you have gotten used to what 99% can and can't do and therefore just dismiss many solutions without thinking since you know the 99% can't do those thing." Yeah I wonder why these orgs always find these XXX hours & $$$ costs cost-prohibitive even though the benefits have proven to be substantial in cutting edge tech industries. It's almost like they typically dismiss smarter and more cost-effective technical solutions superior to naive "throw hours and hardware at it" approaches, because 99% can't do those things.
    - 5y ago

by urthor 5y ago

oblio 5y ago

You'll probably get a long list of stuff that includes things such as: OSes, compilers, game engines, distributed process frameworks, etc. Which as you'll note, are infrastructure things. Things which need 1/100th the manpower (at most!) to maintain. While everyone else is busy churning user-facing apps. You don't need to do that stuff, you career can be extremely successful without ever going near those things, especially if you are not interested in them.

urthor 5y ago
That's the thing, I think a lot of those are, if not solved, problems where there's a lot of really well established good solutions that exist. I could, I guess, tackle them. But I've got an appreciation that what's already there is already pretty great. I suppose that's maybe the answer to my question. You don't need to be 1% to solve a problem as well as the solutions that already exist. You need to be the 1% to improve on the current state of the art. Which in a commercial environment, as you can imagine, is exceptionally rare. They are more about finding unsolved problems than iterating on problems already solved by FOSS.
- oblio 5y ago
  For various reasons (usually control, but frequently spun as "needing to be cutting edge"), some groups or companies do reinvent those wheels and in some lucky cases we get progress. For example Google has recreated several of those things and especially in the data storage field they've probably moved things forward. But in general I agree with you, the crushing majority of developers out there won't work on those things and won't ever even need to. It's also a source of not always deserved elitism from the people that work on those things, and I definitely dislike that (there's enough elitism in this field, as-is).
  - urthor 5y ago
    Elitism is driven by ego and self actualization though. I think there's a lot of people who are highly, highly motivated to work on tough problems with competent co-workers. Those emotions are less driven by the domain itself, and more about the fact that those people have succeeded in finding meaningful work with co-workers of appreciable worth. The tricky thing I find is finding that. Far too many regular projects are focused on mobilizing the dumbest blocks of wood towards a goal. Without actually investing in the human beings doing the work. I think once we regain a culture of actually investing in people who aren't in the top 10% in their field, things can change. Currently though we do not have that culture.

themacguffinman 5y ago

Building a high availability system? The downtime (especially scheduled downtime) of many of these "99%" services is shocking compared to the cutting edge consumer services we're used to. The "hug of death" is an affectionate meme that basically describes the remarkably common scenario where a "99%" website goes down at the precise moment at where it's peaking in popularity and profitability. Or how about high performance? Google famously knows that every millisecond of delay strongly impacts user retention and bounce rate, yet you wouldn't know it browsing many "99%" services. Achieving consistently low end-to-end latency is really hard regardless of scale, it's so easy for performance issues to creep in from every corner of your stack and accumulate. These are hard problems and IMO they are rarely solved well by just stapling together some useful crap.

falcolas 5y ago
Those are usually business decisions. Pay an extra $$$ for an outage scenario that occurs once a month, and doesn't materially impact our users? Hell no. Spend XXX developer hours and/or $$$ on hardware to make the site 50% faster, instead of adding new features our users pay for? Hell no. And let's be honest, for touting speed, Google's awfully slow and bloated these days.
- themacguffinman 5y ago
  A hug of death when your product has peak user interest...doesn't materially impact users? Your users don't care about multi-hour or even multi-day downtimes in "99%" industries like...healthcare and banking? What a perfect demonstration of "Those scenarios are rare if you have gotten used to what 99% can and can't do and therefore just dismiss many solutions without thinking since you know the 99% can't do those thing." Yeah I wonder why these orgs always find these XXX hours & $$$ costs cost-prohibitive even though the benefits have proven to be substantial in cutting edge tech industries. It's almost like they typically dismiss smarter and more cost-effective technical solutions superior to naive "throw hours and hardware at it" approaches, because 99% can't do those things.
  - 5y ago