1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Early access

Discussion in 'v21 Pre-Release' started by Max, Mar 3, 2020.

  1. HaroldL

    HaroldL Alibre Super User

    Geez, so many hoops to jump thru to get Alibre to do what it's supposed to do. Oh wait, didn't a lot of these issues start with HOOPS?
    simonb65 likes this.
  2. simonb65

    simonb65 Alibre Super User

    When is the application going to be regression tested? For example, in applications I write, if I change something like using a new library or component, then functionality and behaviour that occurred before the change is tested after the change to ensure it behaves and functions EXACTLY the same. There seems to be many issues and differences between Legacy Display and HOOPS that shouldn't have even left the developers PC/Laptop as the unit testing (Alibre does unit testing ... right?) will identify these issues! Every bug or issue should be added to the unit tests or test plan.

    The user should not see any changes between these two rendering cores other than its performance (faster!) and appearance (shading, antialias, etc).

    It should not be released to customers without these basic tests!

    Now I know v21 has added more functionality, which is very nice, but WHEN are the basic long term issues and issues introduced by HOOPS going to be resolved. HOOPS was the major release feature of v2019 and now we are running Legacy Display because HOOPS has so many BASIC rendering and scaling issues! So what was that years maintenance actually for?

    Guys, we are paying business customers, NOT casual hobby users who paying for the privilege of being alpha/beta testers! In my software development, I have to pay/give incentives a select core of my most valuable and close customers to do that!
  3. Notermans

    Notermans Member

    After nearly a year, the problems are still not resolved.
    the update of V19 was a disappointment
    and that of V21 as well

    a few new futures are good, but that does not outweigh the problems that have not yet been resolved.

    if I deliver a bad product to my customers, they will not pay
    I assume it should be the same with you
    I wish you many successes in solving the problems

    kind regards, Jos Notermans
  4. JST

    JST Alibre Super User

    Well, I am using the HOOPS display system, not legacy. And the issues that I have seen have been BOTH "transient" (they vary with time, and even "come and go" over time), AND they have been associated with specific largish files that were first created several releases ago (as in 2014 type time period).

    I cannot fault the logical argument about development given above. And I have some concerns that the present development model, one of several incremental changes and one major change per year, may be driving development to choose output over complete testing and correction before release. Microsoft has made a successful business model of doing just that, releasing "beta plus" software as a product, but they also have a near-monopoly.

    However, Alibre works, other software is not perfect, and the job is in general do-able with Alibre.
    Last edited by a moderator: Mar 27, 2020
  5. Max

    Max Administrator Staff Member

    Guys - breathe.

    Things I've seen so far here and on another thread with many of the same people:

    1) We do no regression testing (we do lots of regression and unit testing throughout the release cycle, automated and manual)

    2) Alibre Design's display pipeline is rife with bugs (in fact, the bugs you guys describe here are very rarely reported - so rare and typically geometry dependent that they are not the kind of thing a regression test would catch. I think we've seen less than 5 people report the weird see-through aliasing bug JST sees often, worldwide, in over a year...for example...) Does that bug make for a messy screenshot and 10 threads on here? Sure. Is it actually a big problem for more than a minuscule number of people? No. Should we try to figure it out? Yes. Is that bug a super high priority compared to other things? No.

    The display pipeline isn't perfect. It can and should be improved - but it is not unusably bad by any stretch for the vast vast majority of business customers. A lot of people quite like it. The forum is a good way to find an echo chamber on these kinds of things, especially when some folks post the same topic over, and over, and over and assume everyone else must be having the same issue with the same frequency. That said, the large dim bug on some imported files is something that does need to be addressed. I'm not trying to minimize the issues that the few are facing, just asking to keep context in mind before making blanket statements like "Alibre doesn't test anything" or "HOOPS has so many basic rendering and scaling issues".

    3) We must consider Alibre Design to be a hobby product (No)

    4) That implementing things like a part library are "super easy" (it is not trivial, nor is it cheap - it is something we look at on an ongoing basis to try to find a good solution that's workable). Of those commenting on this - how many of you have tried to make or license a multi-standard part library suitable for worldwide use in half a dozen standards? If you have a solution - by all means tell me.

    5) That our development model is one major change per year and several minor ones (we tried to explain in multiple communications and on here that our new model is to deliver a solid release every few months, 3-4 per year, of about the same scope).

    We've heard comments like "HOOPS just looks nice, it's not a feature I care about - must be for hobby users". The entire reason we considered HOOPS to begin with was primarily for performance. Our homegrown DX9 solution didn't scale. Very large, professional assembly view performance was quite bad. We don't want to spend our time writing graphics pipelines. And many of the upcoming display-related features would never be done on a homegrown system. Could they technically be? Yes. Would we have done it? No. Bigger fish. Point is - HOOPS started as a performance improvement for our most demanding business customers, not for model shadows.

    My intent here is not to be contrarian or to minimize what people are saying, but the assumptions about our internal processes and blanket statements are getting a tad out of hand. They are not helpful and more importantly are usually not accurate.

    That said, we're working hard on finalizing the release which should be out quite soon. Stay safe out there.

    PS. the build on this thread is quite old - and we're close enough to the new release that we aren't going to post a new one.
    Last edited: Mar 25, 2020
    Alexander, oldfox, netcp and 6 others like this.
  6. JST

    JST Alibre Super User

    Thanks.... good to know.

    BTW, I do not see it "often", I saw it on the specific files I mentioned. All of them old files created by Alibre in 2014 and 2015 or so. The bug of seeing edges that should be hidden is a little more common, but still not "often.

    I mention them because they were very noticeable. Having done nothing particular to get rid of them, they still went away on their own, which I found very strange, and worthy of mention. And MORE worthy of mention when they re-appeared suddenly (and since then have gone away AGAIN.).

    After decades of product development, I have learned to thoroughly check out odd things like that, because they can be the very first indications of a problem that will be seen later more frequently. So if I see it I will mention it.
    simonb65 likes this.
  7. domcm

    domcm Senior Member


    What I am more curious about is feedback on the bugs with v21 that have been reported on this thread. Have any been reproduced/fixed?
    simonb65 likes this.
  8. Max

    Max Administrator Staff Member

    We've been tracking them. I just heard we also fixed the zebra stripe issue JST was seeing, though it's possible there are multiple reasons it happens.

    The cosmetic thread bugs are irrelevant as we are not including that in v21. The convert to sheet metal "white boxes" should be fixed. The shaded views issue should be fixed. Edge selection now features a thicker edge - not a bug but was a request and it is much nicer than before. Section View coloring in 2D would turn blue in one particular workflow - it should be fixed.

    We haven't been able to address the large dim in imported parts where the unit of the imported item is different than the assembly bug, but will probably do that in a SP.
    simonb65 likes this.
  9. Lew_Merrick

    Lew_Merrick Alibre Super User

    Hi Max -- My "argument" is based on the late-1970's "256 color scheme" defined by the "National Bureau of Printing" tha has certainly "expired." In those days. a "2;s complement color" was a "higly visible color" against any of the "256 defined colors." In that context my "suggestion" has been that a "phantom line" or "hidden line" cylinder be constructed about the thread -- Major Diameter for female threads and Minor Diameter for male threads. Such a "display" would warn users & viewers of the existance of the threads. Also, when a thread is applied to a Model, the "thread definition" (say, .2500-20UNC-2B) would be the "identifier" applied to that Feature.

    Says the guy who would not have to code it -- "It shouldn't be that tough to accomplish." [Just remember that I spent three years learning to prove mathematically the Murphy is an optimist.]
  10. domcm

    domcm Senior Member


    Thanks for the feedback on bugs that have been fixed. Excited about the edge selection enhancement. Looking forward to the update.
  11. NateLiqGrav

    NateLiqGrav Alibre Super User

    Is it too late to get an additional Apply button positioned next the the Close button on the new constraint tool? I keep mousing there. Old habits are hard to break.
  12. Notermans

    Notermans Member

    in the active section view, the disk that is solid appears to be hollow.
    this is the result of the two colored surfaces.

    is a bit like what the cosmetic thread looks like in the active section view assy.JPG

Share This Page