SW Corp, you're regressing again!! Lofts are now screwed up!

SW Corp,

There seems to be a regression with lofts, AGAIN!

http://zxys.com/2003-loft-segments.png

This loft NEVER segmented before, made in SW2001plus and works fine in
SW2003.

Even with the Tangent or Smooth option or GC Tangent option (which it
did NOT have before) will not fix this!!!

Man, every freaking release you guys botch either lofts or sweeps!!!!!

FREAKING JOB SECURITY for you guys, YES!????

....

 

Also, lofts using "Start Tangency Length" have a greater
influence/weight!!!!

IT'S SCREWED UP!!!!

...

 

That is one of the worst things I've seen.

Caveat: I presume your section are continuous, and not made up of a lot of
tiny little segments? And I presume that the preview you show translates
into faces in the final feature, and that it is not just a preview of the
flow lines?
If both of my presumptions are true, that's some really scary shi*. - And
its still pretty darn scary if even the first presumption is false and
everything in your sections is tangent.

To be frank, I loathe when I have to train our guys on lofts and sweeps (and
fills, come to think of it) because one can't simply communicate how these
things work - because those features still , really, don't. I am tasked
with simultaneously arming new guys with dozens of exceptions and special
cases, three ways to finish each task, and some kind of guidance on which
route to take when (not if) they run into a problem.
The mercenary in me might appreciate the freaky fun because it makes me
'important' and keeps me in demand - I somehow always find a way to beat the
system, even when it makes for late nights and bloody knuckles from all that
wood knocking. But darn it, in really honest moments I just wish to be a
designer and hang all of this CAD stuff.

Thanks for sharing. I'll keep an eye out, report it if/when I come across
it, and miracle out some kind of workaround because that is all that we can
do.

 

'some kind of workaround because that is all that we can do.'

really?!!.......dodgy installs...broken functions...
I want a sp1.1 asp

 

Ed,

The surface edge and gc(surface edge) and sketch are all continuous.
After scrutinizing my model, which honestly is/was not that clean
initially (I've clean it up more with the same results).

Later, I tracked it down to a "constraint curve", used in a fill surface
which is later trimmed with a continuous spline and/btw that surface
edge is clean/continuous. But removing the "constraint curve" removes
the segment problem.
Then, after looking at the "constraint curve", it's comb is not bad, a
slight jog, fairly clean but all on one side, nothing suspect imho.
Then, I clean the comb slightly and... reapply the "constraint curve" to
the fill surface and the loft segments go away?
So, it's that "constraint curve" which effects the loft later, but only
in SW2004 sp1?

Everything was/is tangent and it works fine in SW2003 sp5, originated
from SW2001plus.

Yeah, my issues with this program continue to be it's inconsistency per
release and it blows me away that a CAD company gets away with putting
out this stuff every release. Seriously, any other company who did
this would go out of business... which says something about how the
software industry has this barrier of acceptable mistakes.
Otherwise, yeah, it's like every release the user has to apply
maintenance to the model data because of the inconsistency, that is not
productive. Or, become conservative by keeping all the data frozen in
the version it starts with and not upgrading. Seriously, I can to this
day open the majority of my Pro/e files without failing. (regardless
what some pro/e haters say) I can only guess that the rules within
Parasolid and the rules used by SW Corp change too give inconsistent
results per sp?

I'm not one to give up easily but I've been thinking about stopping this
business personally and getting into something else.
One of my goals this year was/is to think about alternative ways to make
a living (and that is because this business is becoming more and more
about money or cheap clients).

Anyhow, I've sent the segmented loft to the VAR and hopefully it will be
fixed or seen as an example where that condition results as such or why?

later..

 

File was sent to SW but if anyone of you want to look at it...

http://zxys.com/swparts/2004_segmented_loft_zxys-3a.zip (320K)

...

 

Paul,

It gets even weirder. I opened the model in 2003 SP5 and did a cntrl-Q, and
it segmented. The note says you cleaned it up in 2003 SP5 ???

I'm emailing you a screen shot

Regards

Mark

 

Hey Mark,

Yeah, got it.
Hmm, 2003sp5!? Do you have verification on?

I just tested with verification on and it still is fine, no segments,
looks the same.

Thanks for looking at it,.. I guess,... now I'm more confused.

later..

 

Mark,

Sorry, about the off-line sp5 question... duh! Yeah, saw your ng
message and responded here..

Man, that is weird!!
I'm going to upload another version, basically the same but a tad more
cleaner.
(originally this file was a test layout part from SW2001plus, never was
that clean.)

And, this one is also fine, no segments, in my SW2003 sp5 version but...

http://zxys.com/swparts/2004_segmented_loft_zxys-3a.zip

... this one whacks out more in SW2004 sp1

Could you see if you still get the problem in SW2003 sp5?

Anyhow, it really should not do this, imho.

Thanks..

 

Paul,

This one seems to be OK. I notice you removed the influence curve from the
fill. I did this on the first model and it still segmented (only two though)

Regards

Mark

 

Paul,

Nice shape! I'm amazed that the final loft worked in the first place. How
did you know that you could loft a shape from an edge to an intersecting
sketch and have it work? I would never have thought to try it, assuming that
it would fail at the corner where the sketch and the edge meet.

Jerry Steiger
Tripod Data Systems

 

Jerry,

Glad you like it, btw,.. here is what the shape evolved into...
http://zxys.com/blnd1.jpg
Somebody may recognize what car body it is similar too?..

It's not uncommon to use the loft that way, imo. How do I know,..
experimentation and I've used it it in similar ways using other tools
using boundaries.
The one problem with that face is the singular condition and offsetting
or shelling that face is sometimes a problem.

...

 

Paul,

I must say...... I have learned how to surface complex shapes just
from
looking at most of the stuff you post. I don't by any means model
complex parts that often, but I must say, I have used many of your
techniques when I have had to do so. Your surfacing models rock
Paul! It's no wonder why your
models have challenged the soidworks software and it's kernal
capabilities only to bring it to it's knees...... you can probably
almost hear the your SW begging you NOT to REBUILD your models!!!! I
will also admit that I participated in the SW Beta testing and even
won an mp3 player for my efforts. However, I probably had a lot more
free time that you at the time. I remember making a comment in the
beta newsgroup saying something quote "I sure hope Paul S. is testing
2004 on his surfacing models because I sure don't have anything as
complex as his stuff." But..... oh well, I can understand your take
on beta testing if your not being supplemented for your hard efforts..
Anyways, I just want to give you a special kudo's for posting some of
the problems with their software. I, and I hope others, hope that SW
only listen to these problems and truly make it better. It's a
win-win it the long run!

By the way..... you wouldn't by any chance be willing to share the
complete surface model of that car body would you????? I must say
that I'm a fan of your work!

 

"It's no wonder why your models have challenged the
soidworks software and it's kernal capabilities only to
bring it to it's knees...... you can probably almost hear
the your SW begging you NOT to REBUILD your models!!!!"

To claim that this fault lies with the Parasolid kernel
is patently ridiculous.

Since you claim that this problem is related to the Parasolid
kernel (and it most certainly is not) can you tell us why
Unigraphics has no such problems ? SolidEdge also does
not have these kinds or problems.

Looking forward to your explanation of how you know that this
is a issue with the Parasolid kernel.

jon

 

I wasn't going to post anything to you at all but.......

jon,
I'm sorry if your childish, misunderstanding mind always reads black
and white. Guys, I am NOT personally putting the sole blame on the SW
kernal for these errors. However, I expect everyone else, with the
exception of jon, to see the humor in my previous post post. I don't
think anyone who isn't an inside programmer can tell us why there are
so many inconsistancies. (But some how, jon knows where they stem
from, and.... I bet he knows how to fix them!)


Seriously, when did computers become widely available to monkeys with
opinions? GOD D*M! Please, if you want to post something
remotely usefull, do the this ng a favor and douse yourself with
gasoline. After you've waited a minute or two for it to soak into
your skin and clothes,
take a lighter and set yourself on fire. Be sure to record this on
some digital media and then post this to the ng. Only then will we
truely understand how seamless, and unified the hellish combination of
gasoline and a lighter!

 

Paul, take a look at file:

ftp://public.ideo.com/blind/2004_segmented_loft_zxys-3b-biasotti.zip

This file will be removed in 72 hours.

I'm still on SP0 but as NG has determined SP1 is not the issue. The
endpoint of your spline in Sketch35 was not merged with your
intersection curve.
The those kind of ripples you see in Fill are caused 95% of the time
by constrain curves that are not properly connected.
As for Loft19, SWX shouldn't behave that way and its in error in
creating the isoparm lines. It can be corrected however by changing
the direction of the loft to use sketch46 and edge1 and picking at
there ends furthest from there converging point at the rear of the
car.

Of course, you can not offset this surface, and should use 3 sided
loft when the final product will be shelled (or thin sheet as in the
case of car body).

Nice model.

Mark

 

"I wasn't going to post anything to you at all but......."

You basically didn't and you certainly didn't answer the
direct questions I asked you. I'm still waiting for those
answers. Perhaps after reading this your answers may
now come easier. :>)

Lets take a look at what you just posted because it contains
more errors.

"Guys, I am NOT personally putting the sole blame on the SW
kernal for these errors."

SolidWorks does not have it's own kernel. Parasolid is the
main modeling kernel used in SolidWorks.

"However, I expect everyone else, with the exception of jon,
to see the humor in my previous post post."

I'm sure plenty of people saw the humor... just not for the
reasons you think.

"I don't think anyone who isn't an inside programmer can
tell us why there are so many inconsistancies."

It doesn't take a rocket scientist to figure out that the
most robust solid modeling kernel on the market
(Parasolid) is not the cause of these problems as they don't
happen in SolidEdge or Unigrahpics. They also don't happen
in other packages that use the Parasolid kernel. Based on
this it's not hard to realize where the problems are. That
would be the layers of code above the Parasolid kernel.
Who writes that code ??? :>)

"But some how, jon knows where they stem from...."

Anyone using some common sense can figure out where they come
from and who is writing the code that is causing these kinds of errors.

"and.... I bet he knows how to fix them!"

Sure do. I've said it for years in this newsgroup.
Implementing ACIS would be a great start. ACIS includes many
surfacing functions that Parasolid does not. It should be
obvious by now that writing code for surfacing is not
SolidWorks Corp. specialty... and that's being more than kind !!!

The most interesting thing about your posts is that you
somehow think that one needs to be a full time CAD expert to
create very pleasing aesthetic shapes. I'd describe this
notion as being akin to being brainwashed.

Instead of hero worship, have you ever considered using a
tool that had a different philosophy so you could easily see
that you could create these kinds of pleasing aesthetic shapes
for yourself ?

jon

 

Hey Mark,

Thanks for looking at it.

Hmm, I do not see a merge constraint in your relations list and do not
see how that is?
I do notice you slightly tweaked the spline though.

So, what I did to check was I removed the splines endpoint constrain,
removed the intersection curve and tweaked the spline for fun... put the
constraints back, same relations as yours,... and again, no merge exist
or is present in the relations list.

So, I did what I suspect you did, I slightly tweak the second to last
points handle, and, walla! Fill6 does not have any turbulence!

Still, it remains, why is there no turbulence in SW2003 sp5 and
SW2001plus?
Why does it suddenly occur in SW2004?

Yeah, it has been sent to SW and GEI and they recognize the isoparm
problem.
I agree, changing or editing most of my lofts and changing order
sometimes fixes this.

Thanks, man!

 

BTW, just by tweaking that last spline point on sketch35, you may also
notice, loft21 may most likely solve properly?

It's just too inconsistent how each version of the software acts
differently.

...

 

BTW to clarify for those who have not used "merge", it is a constraint
which is "STILL" not shown in the relations list.

Merge occurs when two endpoints coincident when dragging, and infer/auto
relations are on, they are assumed to be merged.

But, there is a manual merge, ctrl select two endpoints, you then have a
merge option in the constraint list, but "still" merge does not exist in
the relations list.

Unfortunately, Mark, this is not what changes/fixes the turbulence in
this case. You must have accidentally moved/tweaked the spline when
removing the intersection curve and endpoint constraint.

...