Freeze all development. Fix all crashing bugs

[roambotics_scott]

So I'm told that for this last release, AD addressed a whopping 40% of their CER buckets.

I'm not great at math, but by my county that leaves something in the neighborhood of 60% of known crashing bugs still there.

So my suggestion is simple.

Freeze all new development.
Give all developers access to the CER database.
Fix every crashing bug in there.

Every patch in its own branch and integration test to death as they're merged into master.

Once done, treat every CER report like a loose bolt on a Falcon 9.

I know it's a lot and seems insurmountable but I'd bet it could be done in under a month with all resources focused on it.

That's what «serious about stability» and «empathizing with the users» means.

[CarlM89]

F360 rarely crashes for me. Perhaps once a month. and i use it for hours every day on multiple computers. If you are having regular crashes, perhaps your system has a hardware issue. A failing video card would likely cause crashes.

[roambotics_scott]

Well that's great for you @CarlM89

Per @prabakarm a full 60% of CER buckets (read: observed types of crashes) were unaddressed (read: objectively lots of crashing bugs are in there whether you're personally hitting them or not)

You - and even lots of people - might walk through a minefield unscathed, but that doesn't mean it shouldn't be cleared.

[roambotics_scott]

Also RARELY

Something tells me that those «rare» crashes might be a much bigger deal for you if you had a small business that relied on F360 and needed to get things done.

[CarlM89]

I work for 2 small companies that do rely on F360 for design and CAM and it rarely lets us down. F360 has less bugs than many much more expensive software packages that are not still in rapid development. When I have had problems F360 staff has always provided me with quick support to find a workaround. Generally these issues have always been addressed in the following update. I think their development/bugfixing ratio is ideal.

[roambotics_scott]

Couldn't disagree more but fair enough.

[CarlM89]

Ask me again when I'm having one of 'those days' and I might be singing your tune. Hope your problem is resolved quickly.

[joseangeljimenez]

I am also one of those users suffering several *daily* crashes. My current laptop is a Macbook Air from last year with maximum RAM and SSD. I try to save frequently but even doing so, from time to time I lose some data. 😞

Best regards

[daniel_lyall]

@roambotics_scott for once I have to aggrey with you untill they know what it is, that make the bug`s you and a small amount of user`s are seeing (or just get one interin to work out what computer setup is common to this problem) once they know what the common set up is then they can stop this small yet very bad problem happening.

and yes less work arounds will be needed then too.

keep up the push keep it real and all and any sugestion`s just do it someone has to 

[roambotics_scott]

I do appreciate that, @daniel_lyall

[daniel_lyall]

@roambotics_scott you know they wont do it

[roambotics_scott]

Most probably, but it is a mistake to confuse very, very difficult for impossible and the only guarantee of failure is not to try.

[daniel_lyall]

that`s a interesting saying.

If you dont try you will never know 

[roambotics_scott]

My problem is not that I'm negative, it's that I'm unreasonably optimistic.

[kb9ydn]

I'm going to be the voice of dissent here and say they should *not* do this.  I think it would be better to instead have a beta version and a stable version, where new features are first added to the beta version for public testing before they go into the stable version. 

Managing two versions is more difficult to do but it should improve the product quality significantly while still allowing visibility of new features.

C|

[CarlM89]

I agree kb9ydn, This would be an ideal solution.

[roambotics_scott]

@kb9ydn those aren't remotely exclusive. 

I'm way behind having an unstable / beta stream.

A (temporary) freeze on development gets them to a stable base then, from there, they build putting new features in the beta stream and only rotating things to production when it is rock solid based on considerable testing.

[kb9ydn]

Ok, I just changed my mind.  They DO need to do this.  I just ran into a situation where I have a sketch with one line and one point; and I can't seem to put a dimension between the line and the point.  They are both totally unconstrained and I can't even place a dimension between them. I can create another line and dimension the point to that, but not to the first line.  ???

Whatever happened to basic CAD being table stakes?

C|

[mmossmanSD6L5]

I disagree, I think they have a good mix of feature development and bug fixing.  I rarely experience a program crash with A360, and I would rather see new features come out.

[TrippyLighting]

Not sure this is practical to do in a corporate, distributed software development project but ithe idea definitely has my vote.

No software shall ever crash!

[schengnz]

The last Fusion360 release introduced 2 major bugs that corrupted my design. In effect, I have to stop using Fusion360 for a week and find something else to do while Autodesk technical support figures out a workaround. I have to applaud Autodes's responsive technical support (*relative* to other large software companies), but nevertheless stabilitiy issues like those are real killer for real project. The project I am working on is a test-run for Fusion360. I am absolutely loving the functionality and usability of Fusion360, but would find it hard to recommend for real projects until Fusion360 irons out a few more bugs. In particular I suspect current version of Fusion360 is not taking well to designs that are heavily parametric/user parameter driven.