5 ms·

Because units don't matter anymore ever since the units command line utility was created. I can use watts/hogshead or horsepower/liter just as easily as kilomet

Because units don't matter anymore ever since the units command line utility was created. I can use watts/hogshead or horsepower/liter just as easily as kilometers or miles. In practical terms, for the UI, it's best to use whatever customers are most comfortable with, since you don't want to scare them away over irrelevant issues.

franciscop 11y ago
I'm sure in 1998 there was a similar utility such as units; still the Mars Climate Orbiter disaster [1] and the Gimli Glider [2] happened. By Engineers or technical people, not the general public. I'm sure everyday small misunderstandings/errors happen from the disparity. [1] https://en.m.wikipedia.org/wiki/Mars_Climate_Orbiter https://en.m.wikipedia.org/wiki/Mars_Climate_Orbiter [2] https://en.m.wikipedia.org/wiki/Gimli_Glider https://en.m.wikipedia.org/wiki/Gimli_Glider
- garrettgrimsley 11y ago
  >I'm sure in 1998 there was a similar utility such as units; still the Mars Climate Orbiter disaster The utility did exist then, but a reading of the mishap report[0] for that event makes it clear that NASA was not using it. It looks like someone assigned a constant using the Imperial measure rather than Metric representation of a value. The function that used this constant did not verify the unit of measure, and the constant was likely just a floating point number, not a data type that even stored unit of measure. I haven't read the Gimli report yet, but it's probably the same story: human error. Edit: The Gimli Glider [1] was caused by human error. [0] ftp://ftp.hq.nasa.gov/pub/pao/reports/1999/MCO_report.pdf see page 16 [1] http://aviation-safety.net/database/record.php?id=19830723-0 http://aviation-safety.net/database/record.php?id=19830723-0
  - mpweiher 11y ago
    > probably the same story: human error. Right. Humans are error-prone. That's why you want to remove places where human can make errors, such as unnecessary unit conversions.
    - garrettgrimsley 11y ago
      I agree, and I don't mean to argue against moving to the metric system. I only intended to dispel the notion that software safeguards that would have prevented these incidents were in place and failed. Educating developers about this potential issue and ensuring safeguards are in place to prevent disaster seems like a more readily achievable goal than inducing a national migration to the metric system.
      - Ntrails 11y ago
        Sod national, I'm talking about corporate. If my company mandates that all code must be written in Lisp, then that's what I'll do. If they mandate all monetary values must be stored as USD then that's what I'll do. And if they require that all code must use the metric system - with display options to convert to imperial? Thats. What. I'll. Do. I don't care what the coding guidelines are - I only care that they exist. [within reason]
    - zeveb 11y ago
      In aviation the standard units are traditional units; it's the French units which are non-standard and unnecessary. So the Right Answer is to switch to traditional units everywhere:-)
- sokoloff 11y ago
  The Air Crash Investigations episode on the Gimli Glider is interesting: https://www.youtube.com/watch?v=cf2KCCv-NDo https://www.youtube.com/watch?v=cf2KCCv-NDo
majewsky 11y ago
Fun fact: I recently used units(1) to convert a power of two of something, when it threw a syntax error on me because the power operator (like "2^6") was not known. It turned out that I was accidentally not using the units(1) from my Linux VM, but the one on my Mac's native terminal, which according to the manpage dates back to 1991. Apple has literally not gotten around to updating their units for 25 years.
- coldtea 11y ago
  Actually Apple just uses the BSD-licenced version of units -- and not the GNU units -- which is currently at version 2.x. You can get "gunits" via Homebrew though...
jzwinck 11y ago
As a former draftsman: this is not accurate. You might think that units are convertible and we have plenty of software for that. But it doesn't work that way. If a drawing says to drill 3mm holes and you have imperial drills, what to do? If a drawing says tolerance is 1/64 inch, how do you use your metric digital calipers quickly to verify it? If you have a paper drawing provided to you, which is the reality of life in the field, how do you visualize all the dimensions in your head if they are in a system of units you don't use every day? If you receive a part with 28mm wrench flats, how can you be sure they aren't meant for a 1 1/8 spanner? There are real mass produced parts where the metric female threads can easily be enlarged (or stripped!) to accept imperial male threads. But the converse doesn't work, and even if you possess a tap you probably don't have a metric thread gauge in your imperial shop. It goes on and on. Software won't eat this problem, so we have to.
- sokoloff 11y ago
  Spot on, with one small exception. I suspect every machine shop has thread gauges for Imperial and metric. I'm a backyard hack and I have both, just because this problem is so common and annoying. 28mm is compatible "enough" to 1 1/8, as 14mm and 9/16 are a very common "either one works" interchange for sockets/wrenches.

franciscop 11y ago

I'm sure in 1998 there was a similar utility such as units; still the Mars Climate Orbiter disaster [1] and the Gimli Glider [2] happened. By Engineers or technical people, not the general public. I'm sure everyday small misunderstandings/errors happen from the disparity. [1] https://en.m.wikipedia.org/wiki/Mars_Climate_Orbiter https://en.m.wikipedia.org/wiki/Mars_Climate_Orbiter [2] https://en.m.wikipedia.org/wiki/Gimli_Glider https://en.m.wikipedia.org/wiki/Gimli_Glider

garrettgrimsley 11y ago
>I'm sure in 1998 there was a similar utility such as units; still the Mars Climate Orbiter disaster The utility did exist then, but a reading of the mishap report[0] for that event makes it clear that NASA was not using it. It looks like someone assigned a constant using the Imperial measure rather than Metric representation of a value. The function that used this constant did not verify the unit of measure, and the constant was likely just a floating point number, not a data type that even stored unit of measure. I haven't read the Gimli report yet, but it's probably the same story: human error. Edit: The Gimli Glider [1] was caused by human error. [0] ftp://ftp.hq.nasa.gov/pub/pao/reports/1999/MCO_report.pdf see page 16 [1] http://aviation-safety.net/database/record.php?id=19830723-0 http://aviation-safety.net/database/record.php?id=19830723-0
- mpweiher 11y ago
  > probably the same story: human error. Right. Humans are error-prone. That's why you want to remove places where human can make errors, such as unnecessary unit conversions.
  - garrettgrimsley 11y ago
    I agree, and I don't mean to argue against moving to the metric system. I only intended to dispel the notion that software safeguards that would have prevented these incidents were in place and failed. Educating developers about this potential issue and ensuring safeguards are in place to prevent disaster seems like a more readily achievable goal than inducing a national migration to the metric system.
    - Ntrails 11y ago
      Sod national, I'm talking about corporate. If my company mandates that all code must be written in Lisp, then that's what I'll do. If they mandate all monetary values must be stored as USD then that's what I'll do. And if they require that all code must use the metric system - with display options to convert to imperial? Thats. What. I'll. Do. I don't care what the coding guidelines are - I only care that they exist. [within reason]
  - zeveb 11y ago
    In aviation the standard units are traditional units; it's the French units which are non-standard and unnecessary. So the Right Answer is to switch to traditional units everywhere:-)
sokoloff 11y ago
The Air Crash Investigations episode on the Gimli Glider is interesting: https://www.youtube.com/watch?v=cf2KCCv-NDo https://www.youtube.com/watch?v=cf2KCCv-NDo

majewsky 11y ago

Fun fact: I recently used units(1) to convert a power of two of something, when it threw a syntax error on me because the power operator (like "2^6") was not known. It turned out that I was accidentally not using the units(1) from my Linux VM, but the one on my Mac's native terminal, which according to the manpage dates back to 1991. Apple has literally not gotten around to updating their units for 25 years.

coldtea 11y ago
Actually Apple just uses the BSD-licenced version of units -- and not the GNU units -- which is currently at version 2.x. You can get "gunits" via Homebrew though...

jzwinck 11y ago

As a former draftsman: this is not accurate. You might think that units are convertible and we have plenty of software for that. But it doesn't work that way. If a drawing says to drill 3mm holes and you have imperial drills, what to do? If a drawing says tolerance is 1/64 inch, how do you use your metric digital calipers quickly to verify it? If you have a paper drawing provided to you, which is the reality of life in the field, how do you visualize all the dimensions in your head if they are in a system of units you don't use every day? If you receive a part with 28mm wrench flats, how can you be sure they aren't meant for a 1 1/8 spanner? There are real mass produced parts where the metric female threads can easily be enlarged (or stripped!) to accept imperial male threads. But the converse doesn't work, and even if you possess a tap you probably don't have a metric thread gauge in your imperial shop. It goes on and on. Software won't eat this problem, so we have to.

sokoloff 11y ago
Spot on, with one small exception. I suspect every machine shop has thread gauges for Imperial and metric. I'm a backyard hack and I have both, just because this problem is so common and annoying. 28mm is compatible "enough" to 1 1/8, as 14mm and 9/16 are a very common "either one works" interchange for sockets/wrenches.