Leaderboard

@masao_8 Here is a solution for a simple single selection add and SHIFT-Select to remove. Perhaps this will give you a basis for starting: ;; Function to do a simple Select/Deselect using grread. ;; By PJK - 6/16/2026 (defun pjk-grread-Select (/ done en grl grc grv ss) (if acet-load-expresstools (acet-load-expresstools)) (setq ss (ssadd)) (princ "\nSelect to add objects or SHIFT+Select to remove from selection set: ") (while (not done) (setq grl (grread T 15 2) grc (car grl) grv (cadr grl) ) (cond ((= grc 3) (if (setq en (car (nentselp grv))) (if (acet-sys-shift-down) (progn (if (ssmemb en ss)(ssdel en ss)) (redraw en 4) ) (progn (ssadd en ss) (redraw en 3) ) ) ) ) ((= grc 2) (setq done (if (vl-position grv '(13 32)) T nil)) ) ((= grc 25)(setq done T)) ) ) (if (> (sslength ss) 0) (progn (foreach i (mapcar 'cadr (ssnamex ss))(redraw i 4)) ss ) nil ) )

I have noticed the drop in posts in forums like Lisp, here, Autodesk and Theswamp, I think it's because people are using AI more to write code. Even I try AI now and again when stuck, it is successful some times. The most obvious here is when the AI code does not work help is asked for here. Maybe Admin could comment about number of posts say compared to 1 or 2 years ago ? Or is it that a lot of new users just don't ask for help ? I know I push for process improvement and have often tried to influence people into saving time but the majority just don't care. Any body else want to comment ?

I have mixed feelings about Clippy the AI , at least the one from Google , sure its fast and it can give some relevant answers but you have to double check everything. I am trying to create a dcl editor. Wasn't sure where to start so asked Clippy and it gave me some good ideas on where to begin. But as things got complexer it really started to mess things up. And every time you use it you need to explain every thing all over again because at least the free / unregistered version remembers (saves) nothing . But I learned from my mistakes and started to write a very extensive manual so next time I can upload that so we can better pick up from where we started. My biggest concern is that it creates a black box. You ask something , it gives you something back and you paste and test the code , great , moving on to next part. After a while you end up with a bunch of working code until it doesn't any more and that's where the trouble begins and you notice you have lost your grip on the code because Clippy did it all for you. Like your daddy did all your homework for you and you have to take the exam and ...oops , daddy aint around now is he?

There is no doubt that AI is replacing (and will do so even more in the future) the need to interact with something or someone in order to find ideas and the motivation to pursue them. I don’t know which AI tools are the best for programming (I’ve only used Chatgpt so far), but my experience has been positive in terms of how stimulating it is to have someone to discuss ideas with and refine them while solving problems. This was something that, in the past, could only be found in forums like this one. However, the code suggested by Chatgpt is almost always lengthy and often fails. I suppose that will change over time.

I also think that its AI, especially for new users. I mean AI is good at coding or spotting errors and asking an AI often gets you an answer way faster than starting a forum post. For example here is a chart of stack overflows posts

Hello, I'm back here once again to bring up a very interesting topic: How can AutoCAD (installed on a Windows 10 tablet with an integrated GNSS receiver) be made to display its real-time UTM position (or latitude/longitude), for example through a block? Is this possible? Are there any Lisp routines capable of achieving this?

AI is a mixed bag for me. Copilot can be infuriating at times — it keeps using LISP functions that aren’t supported in AutoLISP, or ACAD‑only functions that don’t work in BCAD. It can wander off into rabbit holes even after you’ve told it not to, and sometimes it changes working code for no reason and breaks it. But with some persistence and back‑and‑forth, it can produce really good functions. I’ve also tried Claude, which is much better in lots of ways — the code tends to run first time. If i run out of credits then I continue on with co-pilot. I used Fable just before it was banned, and that was pretty amazing too. AI still needs a lot of guidance to produce anything worthwhile. It can be overconfident, sometimes it doesn’t listen and it doesn’t always think clearly about the end goal. I’m currently updating a lot of routines I’ve cobbled together over the years, and AI is helping polish them and fix issues I’d hit a wall with. It’s here now, and it’s not going anywhere. Can forums embrace it? I’m not sure. With the amount of knowledge stored on these pages, maybe CADTutor could even build its own AI. No idea how easy or expensive that would be, mind you...

For Al, I always assume it is nothing more than google on a few steroids, not a lot more. For LISPs - like google, if you ask it for snippets where you are stuck it will get something that works, but something more complex not really. Then you have to break it down get the snippets and sew them together. Anything new or novel that isn't out there is won't cope. So for my work it might help with a few LISPs to improve accuracy or efficiency but the actually money maker and drawing stuff, not seen anything close yet. Made me think though, SLW210, have we actually worked out all (or nearly all) the LISPs we need and just need to find them? (apart from the very specific ones). Last comment - RLX, very true, if I pass anything around the company I need to be able to maintain it, an undocumented 'black box' solution still needs time to make it work.

To be fair Soildworks Forum layout/search is trash. I have never started from the forums to find something its usually a link from a search engine to there. The search engines like google are in a fight right now with AI chatbots and in the end content creators/websites are being affected. going to get worse before it gets better. I genially like AI but its a tool not a solution. but most people unfortunately are ok with "good enough". AI companies are claiming 100x efficiency that is impossible. I liken it to a side scroller game like the dino chrome 404. play that 100x speed and see how far you get.

I think the issue is more generational than anything else. In the forums I keep up with, including here, theswamp.org, and MikeHolt.com, I've noticed that most of the core group have been Boomers. As a whole, younger generations have seemed to visit to get help, but not stay to help others. [There are notable exceptions like Lee.] AI is being oversold. I've tried both Claude and ChatGPT for programming help with LISP. Neither has produced code that worked on the first pass for anything more than a trivial case. We're about even on times when I've had to point out syntax errors in AI code and times AI has found syntax errors in my code. AI is a great replacement for most search engines, but it is not all that great at writing code, at least on the free tiers. AI output is only as good as the instructions it is given. Writing good instructions is hard. AI and the dot com bubble have a lot in common. In five years I expect we will have fewer AI companies, mostly paid tiers, and lots of empty data centers.

Posts are one thing, but what about the actual traffic? Like I stated, most fora already have the answers to questions, so often no need to interact and make a post. CADTutor shows most online 9,928 April 2nd 2026, which was probably a lot of AI bots, etc.

Perhaps you could use the undocumented (acet-sys-shift-down) express tools function within the grread loop? Then you would have to manipulate highlighting with (redraw [3/4]) and use (ssadd) and (ssdel) to update the selection set.

Isn't that how CAD works already? you select something either by mouse clicke or window it will be highlighted hold shift to deselect it the same way. I know if you have to many things selected they are no longer highlighted.

I think you would need to test for 'shift' being pressed and the test for a left mouse entity selection. Do a 'princ' on your grread loop to display what you are doing, shift and select something - which should give you what you want to test for.

;;;You can call Master Leemac's code, but do not modify it (defun c:ttt(/ CPB NEW-PT OBJ PT) (c:cb) (setq cpb (entlast)) (setq pt (cdr(assoc 10 (entget cpb)))) (setq obj (vlax-ename->vla-object cpb)) (setq new-pt (vlax-3d-point (getpoint pt "\n Specify the new location"))) (vla-move obj (vlax-3d-point pt) new-pt) (princ) )

Consider the following example - the selection could be automated if the target block/attribute tag is known, or alternatively, it could search every block attribute for the target field expression and perform a replacement: (defun c:test ( / e n x ) (setq n "$(substr,$(getvar,dwgname),1,3)") (while (progn (setvar 'errno 0) (setq e (car (nentsel "\nSelect field: "))) (cond ( (= 7 (getvar 'errno)) (princ "\nMissed, try again.") ) ( (null e) nil ) ( (not (wcmatch (cdr (assoc 0 (setq x (entget e)))) "TEXT,MTEXT,ATTRIB,MULTILEADER,*DIMENSION")) (princ "\nInvalid object selected.") ) ( (not (and (setq x (cdr (assoc 360 x))) (setq x (dictsearch x "ACAD_FIELD")) (setq x (dictsearch (cdr (assoc -1 x)) "TEXT")) (setq x (cdr (assoc 360 x))) (setq x (entget x)) (= "FIELD" (cdr (assoc 0 x))) (= "AcDiesel" (cdr (assoc 1 x))) ) ) (princ "\nThe selected object does not contain a DIESEL field expression.") ) ( (entmod (subst (cons 2 (strcat "\\AcDiesel " n)) (assoc 2 x) (subst (cons 1 n) (assoc 1 (member (assoc 6 x) x)) x) ) ) (command "_.regenall") ) ) ) ) (princ) )

Having a bit of spare time, I thought I would make a hatch pattern similar to the OS marsh block marsh1.pat