comparative table / superimposed status

[GianFi]

Hello everyone,
I am brand new I have read a lot on the forum and I have seen that this problem is more felt by users of cads dedicated to architecture (naturally) rather than by general cad users, you have present the comparative tables or superimposed state where the differences between current state and state of project are highlighted? Yeah? Okay!
for my bad luck I find myself having to make a discreet amount (about 20) even rather complicated and I am a mortal boredom - they are literally making me fall asleep - also because the designs come from different sources and are not exactly well done, so I wondered if you could have a lisp with which, once you overlap the two states, repass plants and prospects so that picking the points of interest with the mouse, you could break everything
I don't know where to start, I gave the question in the meal to bard and chatgpt, they generated similar code, which conceptually seems (to me) even simple but were unable to generate a correct syntax.
as far as I am concerned I am also willing to learn the lisp rather than to continue to quarrel with lines that do not overlap for a few mm, erases, redraws, breaks, changes color, erases and so on until my head falls on the keyboard.
soon g.

 

[GianFi]

I add some extorted code to chatgpt because I understand that in this way my request is too vague.
the code, data two points picked with the mouse, should. -speak all entities that interfere with clicked points (± 0.3 units) then -delete only broken objects and draw the line between the two points.
the maximum I got is
linelkillernothing: draws the line between the points indicated but does not change the interfering entities
linelkillertoomuch: already better, draws the line but deletes all interfering objects

 

[GianFi]

hi, taken a few steps forward, it certainly does not do everything I want but at least it works without errors and forced me to study some autolisp. therefore:
data two points indicated with the mouse the lisp erases the lines that interfere with said points and creates a new line with current attributes.
I would like to implement it (very) but I am not really able.
a couple of consideration on the aisles. chatgpt and bard can write code on the basis of textual indications provided by the user but they commit a lot of errors even trivial and are extremely prolix of sentences that do not explain anything, in practice at every post repeat what the code does, and that ka@ you I just told you what I want you to do, but they were useful to have a track from which to leave, instead they only waste time.
bing of microsoft that I had already tested with satisfaction for the vba simply refuses to process a direct request but if you feed him some code is able to analyze it step by step, and it is prodigal of explanations, I found it very informative. it also proposes solutions even if they often contain errors, it takes a bit of an eye.
In short, I put the result of my conference as a neophyte of autolisp with the aisles, maybe to someone come the way to help me to deepen

 

[Cristallo]

Forgive me
you would like to make a machine a work of discernment between more or less superimposed entities and/or moved, that among them do not have a common base (different layers, different colors, maybe different types of line...) thinking that (a machine - I repeat it) can understand that that line is associated with that parallel to its sx (as I draw, but not with the parallel to dx that is a furniture) but that to a certain point. . .

and you would also want (a machine) to have discernment to recognize if 2 lines belonging to the same representative category (we say) are very lightly moved, but must recognize on its own if that line is a girdle and not the projection of a countertop. . .

Perhaps sooner or later it will be possible (that I know now no) but when this happens you will not complain if you stay home without work. . .

 

[GP.]

I see in your profile that you use gstarcad, the drawing appears express tools does not do exactly what you ask, but could it be enough?

 

[GianFi]

oh welcome gp, the background problem is that the drawings come from different sources, from cad also very old and are not precise motions, drawing appears becomes illegible as well as overkill does not give me the hoped results, so that I have a little set aside. in the lisp I wanted to insert a tolerance value, like that of overkill to be understood, but the aisles have elegantly ignored it.
You say it's too much for my poor means?

 

[Cristallo]

these ai are more advanced than I expected.. .
in all cases my thinking is: artificial intelligence +uman intelligence = 100% intelligence
from this postulate draw the obvious consequences.

 

[GianFi]

hi crystal, sorry I missed your previous post, and what to say yes! - Yes!
Let us say that I would like the properly driven machine to have a little more initiative.
I don't think I'm so far away: the selection the cheat, the new line builds it, the selective cancellation no, that is a problem, at least for me