As a like-/counter-anecdote, I developed semiconductor CAD tools for 10 years, after spending 10 years using them. When they first started being developed with GUIs, GUI meant UI, and its oft-maligned sibling, UX, wasn't a term. In my learnings, Xt (XToolKit) started putting words and code behind the abstract patterns in the late 80's, but our tool usability suffered horribly as more and more (usually nonresizing) Athena widgets were crammed into every goddamn corner of the screen with microscopic b&w pixmaps. Because of the lack of distinction between UI & UX in the tool design process, tools were extremely challenging to navigate with each new feature-rich release.
One of my first tasks as a project manager in 2004 was to introduce the web concepts of UI/UX into what had become essentially commandlines converted to Tcl/Tk (after Xt we went to Tcl/Tk.. ugh).
First challenge was to convince the old timer CAD devs.
Once I was able to explain there the difference between UX and UI, it waslike a light went on over everyone's head: how you use it is different from what it looks like. I know, obvious now, but not 15 years ago. We spent 10 months really driving the new buzzword UX/UI in order to get buy-in for profiling how the top 3 existing CAD tools (formal verification, layout, and timing) were being used via instrumentation and interviews. We then proceeded to completely redesign the GUIs in Qt using a consistent set patterns, icons, and workflows.
Then we had to convince the old timer engineer users.
We put a lot of effort into classes to explain how to migrate, and holy shit did we get yelled at. So much "It worked fine before, why did you change it?!?!?!?" Uhh... because a feature you use 80% of the time required 5x more clicks to get to than a feature you used 20% of the time? FML. It got better, people liked it more on our follow ups months later. [The first product to use the new suite completed in 12 instead of 18 months and I personally believe it was due to the new tools being faster, but I'm biased, and it could have been a variety of factors.]
I agree with your point that it is frustrating as fuck when a UI/UX pattern changes, and it should not be done glibly. But I have also found myself getting angry at having to adapt to a new change that ultimately made me more productive, just because of my own inertia.
/shrugs/
PS. Ironically, as a sad end to this story: the GUI's my team made in the early/mid-2000's eventually bloated after 10 years in almost the same way the original AIX/Sparc GUI's I used in the early 90's did. New coders came on board, and new managers, and they just crammed new buttons into to the tools without thinking about the UX. That was ca 2010 when I left, so I don't know where they at today, but I did have a "the more the things change, the more they stay the same" moment!!
Fascinating! As somebody who uses that kind of software, old timer electrical engineers seem to be very very unforgiving to UI problems and even more unforgiving to any change for better or worse. I learned that while volunteering for UI/UX at horizon-EDA which aims to become a more usable kiCAD¹
I have never seen worse UX/UI than in electrical engineering tools and I worked a lot with 3D software. They are completely inconsistent with other software, often even with themselves. It often resembles the heating room of a 500 year old building were everybody added things but nobody deared to clean up the things that were already there.
My suffering as a user of such tools motivated me to change things for the better. I never got the idea behind resisting change in UI/UX. It seems to be rooted in the believe that change in UX always means change for the worse and never for good. Which is weird, because even someone like a carpenter is very much interested in the usability of their own tools.
Maybe the problem is that each change in the software means they have to adapt and this demands a certain adaptability, or a will to stay on top in a changing world. It certainly costs energy to do that.
> So much "It worked fine before, why did you change it?!?!?!?" Uhh... because a feature you use 80% of the time required 5x more clicks to get to than a feature you used 20% of the time? FML.
Well, they're not wrong because every other CAD tool simply rearranged the UI for no reason.
In addition, the problem with semiconductor CAD tools is that any feature which isn't used by everybody is effectively broken because it has zero users, to an engineering approximation. I wish my CAD tools had a "CAUTION: this feature was used N times in the last 180 days by all users of the tool where N < 10. Expect bugs".
Although, in terms of UX I've never understood why CAD tools don't use Pie menus--games adopted them eons ago. (Fusion 360 is the exception, and it's a wonderful breath of fresh air).
I'm really curious where you worked now, as I don't remember any of my VLSI CAD tools getting an effective UI makeover (and we used a lot of them). Although I'm pretty sure we skipped most Mentor tools.
>"CAUTION: this feature was used N times in the last 180 days by all users of the tool where N < 10. Expect bugs"
I doubt it would work, power users turn off tracking. Presumably you can't trust that the tracking will return only what they say it will (and not your internet history) but you can trust that opting to turn it off will turn it off.
One of my first tasks as a project manager in 2004 was to introduce the web concepts of UI/UX into what had become essentially commandlines converted to Tcl/Tk (after Xt we went to Tcl/Tk.. ugh).
First challenge was to convince the old timer CAD devs.
Once I was able to explain there the difference between UX and UI, it waslike a light went on over everyone's head: how you use it is different from what it looks like. I know, obvious now, but not 15 years ago. We spent 10 months really driving the new buzzword UX/UI in order to get buy-in for profiling how the top 3 existing CAD tools (formal verification, layout, and timing) were being used via instrumentation and interviews. We then proceeded to completely redesign the GUIs in Qt using a consistent set patterns, icons, and workflows.
Then we had to convince the old timer engineer users.
We put a lot of effort into classes to explain how to migrate, and holy shit did we get yelled at. So much "It worked fine before, why did you change it?!?!?!?" Uhh... because a feature you use 80% of the time required 5x more clicks to get to than a feature you used 20% of the time? FML. It got better, people liked it more on our follow ups months later. [The first product to use the new suite completed in 12 instead of 18 months and I personally believe it was due to the new tools being faster, but I'm biased, and it could have been a variety of factors.]
I agree with your point that it is frustrating as fuck when a UI/UX pattern changes, and it should not be done glibly. But I have also found myself getting angry at having to adapt to a new change that ultimately made me more productive, just because of my own inertia.
/shrugs/
PS. Ironically, as a sad end to this story: the GUI's my team made in the early/mid-2000's eventually bloated after 10 years in almost the same way the original AIX/Sparc GUI's I used in the early 90's did. New coders came on board, and new managers, and they just crammed new buttons into to the tools without thinking about the UX. That was ca 2010 when I left, so I don't know where they at today, but I did have a "the more the things change, the more they stay the same" moment!!