[Discussion] Coincident vs Point on Object

Have some feature requests, feedback, cool stuff to share, or want to know where FreeCAD is going? This is the place.
Forum rules
Be nice to others! Read the FreeCAD code of conduct!
user1234
Veteran
Posts: 3350
Joined: Mon Jul 11, 2016 5:08 pm

Re: [Discussion] Coincident vs Point on Object

Post by user1234 »

obelisk79 wrote: Sat Sep 10, 2022 4:38 pm However, from a users standpoint how is one functionally different. Both attach objects to each other, in once case it is vertex to edge and in the other it is vertex to vertex. However from a user's point of view I fail to see how they are functionally different other than the fact that the user tools won't allow me to use one tool for both situations.
This is not completely true. I am only a user and like the behavior as it is now. It is explicit and have no room for ambiguity. (I and others provided examples in the thread you linked). I often repaired (not only FreeCAD) CAD models and one of the biggest reason (not the biggest) why a model fails is, that a function is ambiguity, often combined with a sort of automatism.

Maybe this is for a beginner a (very) little bit harder, but, at least i think, for the causal (Free)CAD this is a good natural behavior.


GeneFC wrote: Sun Sep 11, 2022 1:58 pm [Rant]
...
[/Rant]
Your php code seems to have issues. Sorry joke could not resist.

But i absolutely agree with your opinion, that the other many places, even in the sketcher, which is a very good WB, which need UI improvement. This topic is one of the last % of the last ‰, if this counts as issue, at least some others and i, do nout count this.


Greetings
user1234
User avatar
obelisk79
Veteran
Posts: 1063
Joined: Thu Sep 24, 2020 9:01 pm

dead topic

Post by obelisk79 »

dead topic
Last edited by obelisk79 on Mon Sep 12, 2022 3:19 pm, edited 2 times in total.
user1234
Veteran
Posts: 3350
Joined: Mon Jul 11, 2016 5:08 pm

Re: [Discussion] Coincident vs Point on Object

Post by user1234 »

obelisk79 wrote: Sun Sep 11, 2022 4:00 pm have asked in my original post the following direct questions which no one has actually paid attention to with their own responses.
So, the questions I pose are these:
Should these tools be considered for replacement by one that covers both situations?
What potential ramifications/downsides would result from such a change?
What reasonable use-case can be made to keep them separated?
I think i have answerd that, not in form in a list, at least point two and three. To point one, maybe yes, but what is the pros/cons of it, which leads to point 2, which is answered, also in the other thread, which you linked before.


Greetings
user1234
User avatar
paddle
Veteran
Posts: 1396
Joined: Mon Feb 03, 2020 4:47 pm

Re: [Discussion] Coincident vs Point on Object

Post by paddle »

I already voiced my opinion on this quite clearly. Which Adrianinsaval already summed up :
adrianinsaval wrote: Sun Sep 11, 2022 4:53 am Can somebody point to any real downside to combining these functions? I agree that for the user they are basically the same, coincidence is a special case of point on object, if a point is an object (and it is) then logically applying "point on object" constraint between points is equivalent to coincidence, can somebody dispute this? The distinction is only on software and on the stubborn mind of those used to this method. Many other CADs treat them as the same constraint too. Besides saving screen real state this change also makes FreeCAD more accessible to new comers, which is also an important thing that should be taken into account. The people commenting here have obviously already learned so they do not care much about that.
Having two separate tools is useless, it clutters the interface, force people to learn two tools, ie two icons and 2 shortcuts. Then I sometime missclick and select the wrong one and get an annoying error.

For new users it makes life simpler. For power user it makes life simpler. It removes one icon from the UI. Who cares that it is conceptually different (which it really isn't.) what good does it make to any user ?

You may say that it's just a little something and that it's not that hard to learn two commands. But FreeCAD has a lot of small things like this, which in the ends results in the 'steep learning curve' or 'aweful UI' that we always hear about about FreeCAD.
User avatar
adrianinsaval
Veteran
Posts: 5541
Joined: Thu Apr 05, 2018 5:15 pm

Re: [Discussion] Coincident vs Point on Object

Post by adrianinsaval »

chrisb wrote: Sun Sep 11, 2022 10:07 am What do you mean by "real"? Your opinion?
I mean real. Is there any constraining that you currently do that you would not be able to do if they were to be combined? Or is at least significantly more difficult? What if it's implemented the same as the current rad/diam command, one context aware command and the two original commands nested so they can be used when necessary, what would be a disadvantage to that?
GeneFC wrote: Sun Sep 11, 2022 3:18 pm They are not the same at all. What if I want to constrain one degree of freedom and I click near a vertex of some sort? How does FreeCAD decide if I want to constrain one degree or two degrees?
that's not a good argument IMO, we have the same miss click problem with tangency or perpendicular or symmetry... why do we have tangency and end point tangency or symmetry around a point and symmetry around a line in the same tool but we can't have coincidence and point on object in the same tool? And how often does anybody click on a vertex when they want to select a line anyways? Seems like an unlikely scenario IMO, it has never happened to me to those others I listed.

And if we're talking about concepts, can anybody contest my statement that coincidence is a special case of point on object? Why should the UI make a distinction if the underlying concept is the same?
User avatar
paddle
Veteran
Posts: 1396
Joined: Mon Feb 03, 2020 4:47 pm

Re: [Discussion] Coincident vs Point on Object

Post by paddle »

We should just run a poll and be done with that subject.
user1234
Veteran
Posts: 3350
Joined: Mon Jul 11, 2016 5:08 pm

Re: [Discussion] Coincident vs Point on Object

Post by user1234 »

adrianinsaval wrote: Sun Sep 11, 2022 9:06 pm Or is at least significantly more difficult?
Very very dull example, i missclicked here. Where it should come a warning or error or no reaction, it makes a invalid corner instead. Ask other here (for example @chrisb), how often this come from new users. When this is the same, this will appear here more often and it is much harder to find, since the symbol is here a good sign, that something is here wrong. When it is solved with an other symbol with the same command, then this all is ambiguous.

Peek 2022-09-11 23-14.gif
Peek 2022-09-11 23-14.gif (188.31 KiB) Viewed 538 times


More see https://forum.freecad.org/viewtopic.php ... 43#p564243


Greetings
user1234
GeneFC
Veteran
Posts: 5373
Joined: Sat Mar 19, 2016 3:36 pm
Location: Punta Gorda, FL

Re: [Discussion] Coincident vs Point on Object

Post by GeneFC »

paddle wrote: Sun Sep 11, 2022 9:14 pm We should just run a poll and be done with that subject.
Polls on the forum tend to be totally useless. Few responses. Not at all representative of the user base.

Gene
User avatar
adrianinsaval
Veteran
Posts: 5541
Joined: Thu Apr 05, 2018 5:15 pm

Re: [Discussion] Coincident vs Point on Object

Post by adrianinsaval »

user1234 wrote: Sun Sep 11, 2022 9:21 pm how often this come from new users.
so they are suing point on object where they should have used coincidence? or what do you mean? I wonder if incorrect usage would really increase if these are combined, I think the opposite might actually happen.
user1234 wrote: Sat Jan 22, 2022 11:51 pm - which sketch below has 3 lines with a coincident and which has 2 lines with point on object?0.png


When all is one constraint (symbol), you can not see any difference, even though both have different meanings and are completely different sketches.
click one side of the line, done, you can see the difference
- point on object can go over the boundaries of the object, because is is based on the untrimmed base object3.gif


If this would be a combined object, this would be completely misleading.
why? the untrimmed base shape of a point is... a point, this is still consistent. (coincidence is special case of point on object x100)
- what is when this happen?1.gif


If this happens, you will not see the difference, because the symbols are the same.
we have substitution for these kind of situation for end point tangency, why wouldn't we be able to have it for coincidence? (that's would be a nice feature idea btw). In general for most or everything that was mentioned, why is this acceptable for tangency/end point tangency, perpendicular/end point perpendicular and symmetry about line/point but not for coincidence/point on object?
chrisb
Veteran
Posts: 53945
Joined: Tue Mar 17, 2015 9:14 am

Re: [Discussion] Coincident vs Point on Object

Post by chrisb »

obelisk79 wrote: Sun Sep 11, 2022 3:08 pm I'm trying to have an engaging discussion about why they cannot or should not be combined as their function is so incredibly similar it begs the question *why?*.
...
This isn't an attack, and I haven't made any demands or personal slights against anyone. I do, however, expect someone who is against the idea, to be able to articulate a rational reason why it should not change that isn't directly related to personal preference or muscle memory
...
You should simply accept that other people think different than you. Let me exaggerate: *why?* don't you propose to throw tangency in the same pot as coincidence? It is also requiring that two objects share a certain place in space. So please can you give me real convincing example *why?* you didn't propose this?

All the arguments you bring here could be just returned in the same way just exchanging the subjetcs.

You see here a lot of people who prefer separate tools for different things. And for these people a coincidence of two points is something completely different than placing a point on a line. Having these separate is extremely clear and extremely easy to explain to newbies. No Ifs, no Whens no dependencies on selections.

You are right, that the notion of "object" includes everything. I see improvement potential though in the name of the point-on-object function, which should be named more precisely as "point-on-edge", because that's what it is. Then the real argument you are looking for could be: a point is not an edge.

This discussion has the same quality as discussing whether Part or PartDesign is better. The answer is finally: it depends on how you think.
A Sketcher Lecture with in-depth information is available in English, auf Deutsch, en français, en español.
Post Reply