|#1 2011-06-21 18:20|
One thing that I find would be useful is to ignore zoom and pan commands when undoing - just undo the drawing/modify commands. - Is there a setting for this?
|#2 2011-06-22 11:01|
Not to my knowledge, and frankly, I think it shouldn't.
Undoing editing commands without displaying the resulting changes (that's what it would come down to) would create fine opportunities for unnoticed data loss - although many applications do it that way, I think the AutoCAD way for once is more correct...
|#3 2011-06-22 15:11|
It's not an option I would use either.
|#4 2011-06-22 15:20|
When it comes to dataloss, I think this is possible, but it depends upon how you operate the program, consider the following.
If I do a bunch of commands then check to make sure what i did was correct and find out that the second command (let's say) that I did didn't have the correct effect, i would like to be able to watch the error in the drawing while undoing (if zooms were omitted) so that I can stop undoing when the error has been fixed. Currently, I have to undo, pan and zoom to check to make sure that I haven't gone too far while undoing, if i still need to go back I need to undo the new zoom and pan commands before I can undo the last drawing operation - this can lead to data loss as well, if you aren't careful you can go back too far. To undo an operation that was incorrect a lot of time can be wasted entering operations that do not have a result on the final output of the drawing.
|#5 2011-06-22 15:52|
I am with Scott on this one and would welcome UNDO skipping zoom and pan changes. The feature could perhaps be made switchable to satisfy users preferring 'Old School' behaviour.
|#6 2011-06-22 16:43|
Perhaps my comment is only partially related...
I often work in model space through locked viewports on a layout tab. If I use the mouse wheel to zoom in or out then each zoom level change has a change to paper space and back to model space.
One crank of the mouse wheel can require 10-20 undo operations. I sure wish these were at least grouped, similar to when working in model space directly.
If ignoring pan and zoom operations with undo then hopefully these implicit space changes would not be on the undo stack either.
|#7 2011-06-22 16:53|
I am also with Scott on this subject. It would save a lot of time, especially in a 3D-workplace. However I would be thinking of an automatic grouping option for all consecutive view-changes and to have them undone in one step. There supposedly is some option like that in Acad (don't konw about Bcad right now) but it does not really work satisfactory.
|#8 2011-06-22 17:06|
I know what you mean Greg, those space changes add a lot of additional strikes to the enter/space/right mouse button to the task at hand...
From my experiences with Autocad, grouping the pan and zoom commands was useful in reducing key strokes, but it still moved you around the drawing when you didn't want to while undoing.
|#9 2011-06-22 20:48|
Still, I don't see another option than pushing view changes onto the undo stack to make sure one actually sees everything that gets undone.
I agree that undo could group consecutive view changes, at the moment this seems to function correctly for wheel zooms (apart from the scenario with the locked viewports, that's definitely annoying), but should probably be extended to include all kind of display commands.
However, the biggest nuisance for me is that the zoom previous command does not group wheel zooms like undo does - don't remember how ACAD handled this, but anyway, it makes no sense to me.
|#10 2011-06-23 02:04|
Just dreamin - what would be nice would be to jump back to say 15 moves ago, confirm or proceed back or forward undoing single operations leaving later undos alone.
I lose more time burning correct work to get at undos than through view changes.
Perhaps I'm dumb, but I think the undo options rely on knowing beforehand what mistakes you will make.
|#11 2011-06-23 09:35|
Such non-linear undo would of course be nice, but I fear it would introduce a tremenduous amount of complexity - and an unstable undo system is probably the last thing we want...
Ideally, Bricsys could open a second, experimental branch, where features like this could be tested by the inclined bunch of us, before getting integrated into trunk.
|#12 2011-06-23 09:44|
Already UNDO does not UNDO everything, even in AutoCAD.
Lisp statements, settings, changes made externally on the hard-disk by the CAD application, drawing changes made by the internal operations (without use of commands) etc still does not come under the purview of UNDO. Perhaps, that is by design. It is just an observation and a point to note.
Given the fact that there are still a bunch of other things that need improvement in Bricscad, I feel that Undo should be left as it is and not tampered with.
|#13 2011-06-23 11:05|
Good points Rakesh, perhaps they illustrate that it's not too hard for Bcad to pick out what it wants to keep or ignore in the undo set. I don't know, I'm not a programmer.
It seems to me that undo already has some intelligence, it just isn't very easy to get at. Maybe the existing options could be extended, undo has had minor issues in the past (not with stability though), but doesn't seem to have problems now.
At least for me, Bcad doesn't have a bunch of things which need fixing. Refinements such as this would be nice embellishments to what is otherwise a very fine program for my purposes.
|#14 2011-06-23 11:17|
If I were an end-user of Bricscad, I would indeed be happy with what Bricscad provides for what it is priced at. No complaints.
But, being a reseller in India, I am shown a bunch of things by users here that they still cannot do easily with Bricscad. Most of them have already been discussed here, Support Requests filed and Bricsys is aware of them. Many will get fixed and implemented in due course.
What I meant is - I would rather that Bricsys put their resources first to fix those problems than to refine the way UNDO works.
|#15 2011-06-23 11:52|
|#16 2011-06-23 13:34|
Well Rakesh, I guess as a dealer you represent the views of a number of people not just one, while your personal view is like mine - pretty darn good.
Of course you would like your clients' requests to be addressed first, but I am one more subscriber whose desires and opinions are worth no more or less than one of your customers'.
If others have requests, I am interested - perhaps there is a better way of working I hadn't thought of. If they find problems I am also interested - perhaps I can avoid, confirm or reject them.
I think the discussion is good, I don't like to see it shut down 'early'.
|#17 2011-06-23 14:09|
I agree with you. Tomorrow, you will see on my blog a comprehensive list of all the wish-lists and requests we have heard and seen from users in India. And these are stuff that I think are good to have as well. Your comments and feedback would be most welcome.
I do not blindlly support any request the user makes about Bricscad. In fact, some have asked me why there is no Visual Lisp editor or the ribbon interface or palettes in Bricscad. I think these are not imporant, in my view. I have not filed a support request for these items. But what really needs improving are those that affect day-to-day CAD tasks.
|#18 2011-06-23 15:04|
At the moment we bundle realtime operations when undoing. If you perform a sequence of realtime zoom , realtime pan and realtime rotate operations, they will be undone in a single step. So we do a form of lumping, but the sequence of operations is always kept intact.
|#19 2011-06-23 15:44|
"If you perform a sequence of realtime zoom , realtime pan and realtime rotate operations, they will be undone in a single step."
This is true for the realtime rotate command '_rtrotf but not for '_rtrot, at least in my case.
|#20 2011-06-23 16:12|
@Alexander: The realtime zoom operations are not bundled when working though a locked viewport. See SR 28765 for more details.
|#21 2011-06-23 16:23|
I have to agree with Rakesh's comments - this is an item that could/should (your choice, you know mine) go onto a wish list. The priority of the Bricscad team should be squishing the bugs - which they do - so far I am impressed with the stability and speed (<- wow) of the program - the important stuff is light years ahead of the rivals!
A bit of background, I have been a long time AutoCAD user (since the R14 days) and CAD manager for multiple firms; "not undoing zooming or panning" has been something that consistently comes up with every employer and I would guesstimate about 70% of the CAD users that I have dealt with - At the moment I manage a small group of AutoCAD users and run Bricscad at home for my personal work.
|#22 2011-07-04 22:04|
An update: I ended up submitting a feature/support request for this (ignoring navigation while undoing), I was told that seperating navigation from commands the effect entities is more difficult it appears. Having dabbled in software myself, I have a question - if navigation commands can be seperated from drawing commands and lumped together, why can't they be ignored? Don't take this post wrong, I am not bashing anybody (or anything), I'm just curious. The new additions to the changes in how undo handles nav. commands in the new version is a huge improvement! At the end of the day, great software and it is definitley the right way to go for me.