4 ms·

This is tough. In many cases I think that people identify as superficially "worthless", I think there's a justifiable reason for those jobs even if it takes a

by bane 8y ago

This is tough. In many cases I think that people identify as superficially "worthless", I think there's a justifiable reason for those jobs even if it takes a little bit of digging to understand why it exists. For example, I know of a few jobs right now that a person does that, if it disappeared tomorrow, would have no immediate measurable impact on the company those jobs exist within. But wait a year or two and those jobs turn out to generate long-term cost savings to the company. But in other cases, jobs that seem to be superficially necessary turn out to not be needed at all. I just said goodbye to a colleague last week and after they left we went through their entire list of tasks and responsibilities -- things that kept an intelligent, highly educated, adult human fully occupied for more than 40 hours a week and paid well into six figures -- and entirely eliminated or rolled over all of those things with no specific impact to our work. The scary part was that much of it was work producing material that was highly demanded of by one of our customers -- the bulk of those 40+ man hours were reports and other deliverable. We simply told the customer that they wouldn't be getting some of that stuff anymore or it would take a different shape to satisfy the need and they were ultimately fine with it. What really concerns me are those jobs that need filling and nobody knows they exist and there's nobody doing them. Things that would eliminate waste, consolidate work, or expand business, but some collective blind spot prevents those positions from being realized. The classic examples in software are good QA people or in many small companies, good sales and business development people. If this still doesn't make sense to this tech crowd. I'll pose this, in the 90s the revolution was "making software useful to people by making the functions of the software discoverable in a well designed GUI". How many of us sit in front of entire screen fulls of discovery-free GUI-less command-lines all day typing out things that took thousands of hours to master and would be bulk eliminated if somebody just put a nice usable GUI in front of it?

theshadowknows 8y ago
Qa people! I would easily give a finger, maybe not a whole arm, but a full finger to get a single, solid qa person on my team. We execute really well, but I don’t have the eye for detail and my team doesn’t have the skill set. But qa is “non revenue generating” and so we don’t hire for it. Even though we lose tons in terms of man hours and productivity due to errors that a solid qa person would help us catch early on.
TeMPOraL 8y ago
I agree with your overall point, but I don't think the GUI example helps your case here. > How many of us sit in front of entire screen fulls of discovery-free GUI-less command-lines all day typing out things that took thousands of hours to master and would be bulk eliminated if somebody just put a nice usable GUI in front of it? We sit in front of "discovery-free GUI-less command lines" to do stuff that is either impossible to be fully and properly captured by a pointing-device operated GUI, or would become orders of magnitude less efficient if operated through such GUI. The discoverability problem is overblown, IMO - the actual problem is that we've trained people to no longer feel expected to learn a tool before using it. The example I usually give is this: no one in their right mind expects to be able to enter a car for the first time in their life and be allowed to drive on public roads. It is expected of them to go through a couple-dozen long training course and learn a bit of theory. People don't complain about that, because there's a social expectation that you need training. Compare that with people whining that a program is "unintuitive" because it requires you to spend 5 minutes in a tutorial to acquire basic operational competency. Compare that to video games, where again there is an expectation to learn, where people don't complain about tutorials. The only way to make software that can be mastered in 10 seconds from first exposure is to make it have flat learning curve - that is, you can't use it for anything more powerful than what you can learn in those 10 seconds. This approach gives you simple toys, not actual tools. As for discoverability of CLI tools, skimming its manual and looking at usage examples is equivalent to taking a GUI program and quickly skimming through all its menus and buttons. The whole problem here is purely of individual emotions. It's as if some people were simply afraid of reading.
- wilsonnb2 8y ago
  For the record, people absolutely complain about tutorials in video games. The best tutorials aren't tutorials at all. Instead, the game gradually teaches you how to play without explicitly teaching you how to play. Obviously, that task is extremely difficult so a lot of games instead opt for the tutorial.
- bane 8y ago
  > or would become orders of magnitude less efficient if operated through such GUI. I don't disagree. The reason for the re-rise of the CLI is that it's just faster to get stuff executed and easier to script and coordinate things. But there's also a great many CLI tools that don't have obvious command-line uses, or the docs are poor and users of those tools just "know" how to use them because of many hours spent learning them. But new and infrequent users spend lots of time reading docs or rereading them, or looking up examples on the internet and in many cases those uses would be immediately obvious with some radio buttons and a couple buttons. There's also the cases of unbreaking things or unexpected results because of typos. GUIs make the trade-off of learn-ability with operational efficiency and frankly, many of the tools we use don't benefit from being able to type them out quickly and would benefit from just having some buttons to click with all the options specified. Also don't think I'm talking only about mousable GUIs. There's plenty of very good examples of keyboardable GUIs with very fast and efficient use-cases that also maintain good discover-ability.
draw_down 8y ago
Sounds like that person had it made to me. Well paid, work that was ostensibly in demand but nothing bad would happen if it wasn’t received... not bad. I suppose it’s “scary” if you’re incentivized to care about the efficiency of the company, which almost nobody in corporate America is.

by bane 8y ago

theshadowknows 8y ago

Qa people! I would easily give a finger, maybe not a whole arm, but a full finger to get a single, solid qa person on my team. We execute really well, but I don’t have the eye for detail and my team doesn’t have the skill set. But qa is “non revenue generating” and so we don’t hire for it. Even though we lose tons in terms of man hours and productivity due to errors that a solid qa person would help us catch early on.

TeMPOraL 8y ago

I agree with your overall point, but I don't think the GUI example helps your case here. > How many of us sit in front of entire screen fulls of discovery-free GUI-less command-lines all day typing out things that took thousands of hours to master and would be bulk eliminated if somebody just put a nice usable GUI in front of it? We sit in front of "discovery-free GUI-less command lines" to do stuff that is either impossible to be fully and properly captured by a pointing-device operated GUI, or would become orders of magnitude less efficient if operated through such GUI. The discoverability problem is overblown, IMO - the actual problem is that we've trained people to no longer feel expected to learn a tool before using it. The example I usually give is this: no one in their right mind expects to be able to enter a car for the first time in their life and be allowed to drive on public roads. It is expected of them to go through a couple-dozen long training course and learn a bit of theory. People don't complain about that, because there's a social expectation that you need training. Compare that with people whining that a program is "unintuitive" because it requires you to spend 5 minutes in a tutorial to acquire basic operational competency. Compare that to video games, where again there is an expectation to learn, where people don't complain about tutorials. The only way to make software that can be mastered in 10 seconds from first exposure is to make it have flat learning curve - that is, you can't use it for anything more powerful than what you can learn in those 10 seconds. This approach gives you simple toys, not actual tools. As for discoverability of CLI tools, skimming its manual and looking at usage examples is equivalent to taking a GUI program and quickly skimming through all its menus and buttons. The whole problem here is purely of individual emotions. It's as if some people were simply afraid of reading.

wilsonnb2 8y ago
For the record, people absolutely complain about tutorials in video games. The best tutorials aren't tutorials at all. Instead, the game gradually teaches you how to play without explicitly teaching you how to play. Obviously, that task is extremely difficult so a lot of games instead opt for the tutorial.
bane 8y ago
> or would become orders of magnitude less efficient if operated through such GUI. I don't disagree. The reason for the re-rise of the CLI is that it's just faster to get stuff executed and easier to script and coordinate things. But there's also a great many CLI tools that don't have obvious command-line uses, or the docs are poor and users of those tools just "know" how to use them because of many hours spent learning them. But new and infrequent users spend lots of time reading docs or rereading them, or looking up examples on the internet and in many cases those uses would be immediately obvious with some radio buttons and a couple buttons. There's also the cases of unbreaking things or unexpected results because of typos. GUIs make the trade-off of learn-ability with operational efficiency and frankly, many of the tools we use don't benefit from being able to type them out quickly and would benefit from just having some buttons to click with all the options specified. Also don't think I'm talking only about mousable GUIs. There's plenty of very good examples of keyboardable GUIs with very fast and efficient use-cases that also maintain good discover-ability.

draw_down 8y ago

Sounds like that person had it made to me. Well paid, work that was ostensibly in demand but nothing bad would happen if it wasn’t received... not bad. I suppose it’s “scary” if you’re incentivized to care about the efficiency of the company, which almost nobody in corporate America is.