| |
So, I made this tutorial about being a reviewer and review parts in the Parts
Tracker:
http://www.l3go.bugge.com/articles/Reviewing-parts-tutorial.shtml
Eventually it’ll end up in the article section of Ldraw, but right now I’d just
like to hear some feedback/suggestions from the general community, especially
concerning illustrations.
It’s basically a major upgrade of this thread:
http://news.lugnet.com/cad/?n=11914 (wow that was a long time ago!).
It is supposed to be a basic introduction about reviewing, so more advanced
stuff about the correct use of submodels, patterned parts, primitives etc. - and
how to review these is not included: the tutorial is already long enough, and I
don’t know enough about these aspects.
I hope somebody else will be inspired to make a tutorial that takes reviewing
the next step from mine (I’d like to read it!). And of course - that a lot of
people will decide to become reviewers or parts authors - it’s really not that
difficult and with the ever increasing complexity and number of new parts, we
need you to participate!
Cheers NB
|
|
| |
Niels Bugge wrote:
> So, I made this tutorial about being a reviewer and review parts in the Parts
> Tracker:
>
> <http://www.l3go.bugge.com/articles/Reviewing-parts-tutorial.shtml>
>
> {Eventually it’ll end up in the article section of Ldraw, but right now I’d just
> like to hear some feedback/suggestions from the general community, especially
> concerning illustrations.}
A couple of other points that may be worth mentioning:
1) LDView can do BFC checking now (I use this all the time)
2) LDDP has a "fix error" option for most errors L3P can find, including
bowtie quads.
Other than that it looks good.
Thank you
--
Dean Earley, Dee (dean@earlsoft.co.uk)
irc: irc://irc.blitzed.org/
web: http://personal.earlsoft.co.uk
phone: +44 (0)780 8369596
|
|
| |
In lugnet.cad, Dean Earley wrote:
> A couple of other points that may be worth mentioning:
> 1) LDView can do BFC checking now (I use this all the time)
You can still only look at the files in ldview right? (it doesn't have any extra
features compared to lab?)
> 2) LDDP has a "fix error" option for most errors L3P can find, including
> bowtie quads.
Cool, thanks! I'll add a note about that (and use it heavily :D )
Cheers NB
|
|
| |
> > A couple of other points that may be worth mentioning:
> > 1) LDView can do BFC checking now (I use this all the time)
>
> You can still only look at the files in ldview right? (it doesn't have any extra
> features compared to lab?)
Yeah, its just a viewer, but looks a lot nicer.
I never got on with L3Lab, not entirely sure why.
> > 2) LDDP has a "fix error" option for most errors L3P can find, including
> > bowtie quads.
>
> Cool, thanks! I'll add a note about that (and use it heavily :D )
IIRC, it can fix all but the non coplanar quads.
--
Dean Earley, Dee (dean@earlsoft.co.uk)
irc: irc://irc.blitzed.org/
web: http://personal.earlsoft.co.uk
phone: +44 (0)780 8369596
|
|
| |
In lugnet.cad, Dean Earley wrote:
> Niels Bugge wrote:
> > So, I made this tutorial about being a reviewer and review parts in the Parts
> > Tracker:
> >
> > <http://www.l3go.bugge.com/articles/Reviewing-parts-tutorial.shtml>
> >
> > {Eventually it’ll end up in the article section of Ldraw, but right now I’d just
> > like to hear some feedback/suggestions from the general community, especially
> > concerning illustrations.}
>
> A couple of other points that may be worth mentioning:
> 1) LDView can do BFC checking now (I use this all the time)
it will not work on all graphic cards corretly. at least it doesn't on my ATI
Radeon 9000. trinagles with a wrong winding will always show up in correct
green.
> 2) LDDP has a "fix error" option for most errors L3P can find, including
> bowtie quads.
>
> Other than that it looks good.
>
> Thank you
|
|
| |
In lugnet.cad, Niels Bugge wrote:
> So, I made this tutorial about being a reviewer and review parts in the Parts
> Tracker:
you won't believe it but I just learnt a few new things.
> 2) Check for BFC problems and gaps via L3Lab (mark them via MLCAD and fix them via LDDP)
no marking needed:
* open the part in LDDP
* select "Process > External programs > Polling > Poll to L3Lab & LDView" (but
it's faster via the icon in the toolbar)
* select "Process > External programs > Polling > Poll to selected line only"
(but it's faster via the icon in the toolbar)
* select "Process > External programs > Polling > Poll every 1 sec" (but it's
faster via the icon in the toolbar)
* launch L3Lab selecting " the toolbar "Process > External programs > L3Lab"
(but it's ...) you an also launch LDView if your video card supports LDView's
BFCing correctly.
* select "Test > mytest6" in L3Lab
* arrange LDDP and L3Lab windows to see them properly
* place cursor in first line in LDDP
* move down with the arrow key. L3Lab will draw just to this line
* in case you encounter a red triangle, quad select "tools > Reverse Winding"
(but its faster via ...) actually the fastest way is just using the undocumented
F12-key and the arrow-down-key
> non coplanar quads ... can be particularly nasty to get rid of
LDView checks for this errors and offers you a possible split into triangles. be
aware: when you copy the error to the clipboard LDView will round the figures.
as far as I know travis is working on this.
> (for instance if you find a part that has had a hold vote for a very long time without anyone reacting).
we have a new policy for held files:
http://news.lugnet.com/cad/?n=13886
bye, w.
|
|
| |
In lugnet.cad, Willy Tschager wrote:
> > 2) Check for BFC problems and gaps via L3Lab (mark them via MLCAD and fix them via LDDP)
>
> no marking needed:
>
> * open the part in LDDP
> * select "Process > External programs > Polling > Poll to L3Lab & LDView" (but
> it's faster via the icon in the toolbar)
> * select "Process > External programs > Polling > Poll to selected line only"
> (but it's faster via the icon in the toolbar)
> * select "Process > External programs > Polling > Poll every 1 sec" (but it's
> faster via the icon in the toolbar)
> * launch L3Lab selecting " the toolbar "Process > External programs > L3Lab"
> (but it's ...) you an also launch LDView if your video card supports LDView's
> BFCing correctly.
> * select "Test > mytest6" in L3Lab
> * arrange LDDP and L3Lab windows to see them properly
> * place cursor in first line in LDDP
> * move down with the arrow key. L3Lab will draw just to this line
> * in case you encounter a red triangle, quad select "tools > Reverse Winding"
> (but its faster via ...) actually the fastest way is just using the undocumented
> F12-key and the arrow-down-key
Man, that's seriously smart! thanks a lot, I'll add that to the tutorial when I
get some time to look at it again.
> > non coplanar quads ... can be particularly nasty to get rid of
>
> LDView checks for this errors and offers you a possible split into triangles. be
> aware: when you copy the error to the clipboard LDView will round the figures.
> as far as I know travis is working on this.
>
> > (for instance if you find a part that has had a hold vote for a very long time without anyone reacting).
>
> we have a new policy for held files:
>
> http://news.lugnet.com/cad/?n=13886
I'll check that out, thx
> you won't believe it but I just learnt a few new things.
I'm just amazed at the tips and tricks that is beginning to crop up in this
thread with people pouring out their experience - I'm only happy if I'm able to
give something back to the more experienced reviewers ;-)
Cheers NB
|
|
| |
In lugnet.cad, Willy Tschager wrote:
> LDView checks for this errors and offers you a possible split into triangles. be
> aware: when you copy the error to the clipboard LDView will round the figures.
> as far as I know travis is working on this.
I will try to have this fixed in the next release (I haven't implemented the fix
yet). Also, the two triangles that LDView shows you are the ones that it chose
to split along. There will always be two ways to split a quad into two
triangles, and there's no way for LDView to know which one to use (unless the
quad is also concave, which is also bad), so LDView just splits arbitrarily.
The crease that is introduced by the split will be between the two chosen
triangles. If you want the crease to go the other way, you'll have to come up
with the new triangles yourself. Having said that, bent quads are often only
bent a little bit, in which case it likely doesn't matter which way the crease
goes.
--Travis
|
|
| |
In lugnet.cad, Willy Tschager wrote:
> > A couple of other points that may be worth mentioning:
> > 1) LDView can do BFC checking now (I use this all the time)
>
> it will not work on all graphic cards corretly. at least it doesn't on my ATI
> Radeon 9000. trinagles with a wrong winding will always show up in correct
> green.
Just out of curiosity, have you tried using the latest ATI drivers? Or is it a
notebook that the manufacturer hasn't posted updated drivers for? (I know you
emailed me about this a long time ago, but can't remember the specifics.)
Since the feature was originally developed on an ATI card (9700 pro), I know it
can work on ATI cards. The red/green back/front faces feature uses really basic
OpenGL functionality, but ATI historically has had horrendous OpenGL drivers.
They had a lighting bug that affected LDView any time subdued lighting was
turned on with specular also enabled. They fixed that within the last six
months. And that was a bug in basic OpenGL lighting functionality that's been
in OpenGL since it's very beginning.
--Travis
|
|
| |
Niels,
Thanks for putting this together. It is good to encourage new people to become
reviewers, and to reassure them that it isn’t that hard to do.
Now, some of my personal observances:
IMO, the most useful thing for a new reviewer to understand are that there are
three main things to be looking for when reviewing a part
1) Is this LDraw part a good representation of the real LEGO part? (is it the
right size, shape, etc.)
2) Does it conform to similar parts in the library? (does it have the right
origin & orientation, does it make good use of primitives, is it named well)
and
3) Does it have any technical issues that will cause problems for renders?
(stuff like BFC, bowtie quads, non-coplanar quads)
These are covered under the “What should I look for when I review a part?” item
on the Reviewer FAQ http://www.ldraw.org/library/tracker/ref/reviewfaq/, but it
is worth emphasizing.
The second most important thing to understand is that there is a list of things
that are NOT acceptable for a hold vote under any circumstances. These include:
- File not BFC’ed (which is *optional*, not mandatory)
- problems with the title (PT admins can easily fix this, just “novote” it with a comment)
- problems with part number (again, admins can fix)
- KEYWORDS or CATEGORIES you don’t like (again, admins can fix)
- using overlapping ring primitives of the same color (this is an accepted practice)
- the orientation of stud logos (if you feel strongly about it, feel free to submit a comment via a “novote”, but this is NOT an acceptable hold issue)
- items on this page: http://www.ldraw.org/library/tracker/ref/l3pmsg/ which call for a “warn” (i.e. novote comment) instead of a “hold”
- problems visible in POV-Ray that are not visible in LDraw tools. (while rendering well in POV-Ray is nice, L3P isn’t perfect, and sometimes there are gaps. If the problem doesn’t show in LDView, MLCad, etc, this is not an acceptable reason to hold)
there are probably others, but these are the common ones I see. If the parts
admins have any corrections for this list, please respond, but I think it’s
accurate.
The third thing I wish reviewers would keep in mind is that we don’t need to be
prefectionists. You may see a missing edge line when zoomed in 1000 times, but
which would never be visible when viewed at normal size, or when part of a 500
piece model. Is it really necessary to hold for that, especially when the file
was first uploaded in 2002, and has been ignored by reviewers ever since? I
feel that there are a lot of frivolous holds on acceptable parts in the Tracker.
end rant
Anyway Niels, as far as the actual text you wrote, the only problem I have with
it (besides POV-Ray & stud logos), is that a lot of stuff you wrote regarding
bowties, non-coplanar quads, & fixing BFC really apply to authors, not to
reviewers. All the reviewers really need to know regarding l3p errors is to
compare errors found to this page:
http://www.ldraw.org/library/tracker/ref/l3pmsg/ and hold/warn accordingly.
The explanations you give would be good to have in a separate article such as
“What do those L3P -check errors really mean?”
Thanks,
Andy Westrate
|
|
| |
In lugnet.cad, Travis Cobbs wrote:
> In lugnet.cad, Willy Tschager wrote:
> > > A couple of other points that may be worth mentioning:
> > > 1) LDView can do BFC checking now (I use this all the time)
> >
> > it will not work on all graphic cards corretly. at least it doesn't on my ATI
> > Radeon 9000. trinagles with a wrong winding will always show up in correct
> > green.
>
> Just out of curiosity, have you tried using the latest ATI drivers? Or is it a
> notebook that the manufacturer hasn't posted updated drivers for? (I know you
> emailed me about this a long time ago, but can't remember the specifics.)
the latter :-( the driver is two years old and was a one shot for those notebook
series.
w.
|
|
| |
In lugnet.cad, Andrew Westrate wrote:
| |
IMO, the most useful thing for a new reviewer to understand are that there
are three main things to be looking for when reviewing a part
1) Is this LDraw part a good representation of the real LEGO part? (is it the
right size, shape, etc.)
2) Does it conform to similar parts in the library? (does it have the right
origin & orientation, does it make good use of primitives, is it named well)
|
Good points, I think I’ll try to separate these in into a separate section early
in the tutorial.
| |
3) Does it have any technical issues that will cause problems for renders?
(stuff like BFC, bowtie quads, non-coplanar quads)
These are covered under the “What should I look for when I review a part?”
item on the Reviewer FAQ http://www.ldraw.org/library/tracker/ref/reviewfaq/,
but it is worth emphasizing.
|
This tutorial was more or less meant to be an example on how you can review a
part (basically the way I do it, since that’s what I know). I try to include
pieces of information when nessecary (that may be scattered in confusing
compilations around Ldraw or Lugnet).
| |
The second most important thing to understand is that there is a list of
things that are NOT acceptable for a hold vote under any circumstances.
|
Well, I don’t recall encouraging to place “illegal” hold votes in the tutorial -
yes, I write about BFC’ing and studlogos, but only as “best practice” and
associated with novotes...
| |
- File not BFC’ed (which is *optional*, not mandatory)
- problems with the title (PT admins can easily fix this, just “novote” it with a comment)
- problems with part number (again, admins can fix)
- KEYWORDS or CATEGORIES you don’t like (again, admins can fix)
- using overlapping ring primitives of the same color (this is an accepted practice)
- the orientation of stud logos (if you feel strongly about it, feel free to submit a comment via a “novote”, but this is NOT an acceptable hold issue)
- items on this page: http://www.ldraw.org/library/tracker/ref/l3pmsg/ which call for a “warn” (i.e. novote comment) instead of a “hold”
- problems visible in POV-Ray that are not visible in LDraw tools. (while rendering well in POV-Ray is nice, L3P isn’t perfect, and sometimes there are gaps. If the problem doesn’t show in LDView, MLCad, etc, this is not an acceptable reason to hold)
|
It would be really nice to have these issues+responses compiled in a single
document on Ldraw: Something I could link to without drowning the simple baking
recipe-style tutrial I’ve tried to make (like as: “...just start doing this, and
you’ll pick up the rest as you go along”)
| |
The third thing I wish reviewers would keep in mind is that we don’t need to
be prefectionists. You may see a missing edge line when zoomed in 1000
times, but which would never be visible when viewed at normal size, or when
part of a 500 piece model. Is it really necessary to hold for that,
especially when the file was first uploaded in 2002, and has been ignored by
reviewers ever since? I feel that there are a lot of frivolous holds on
acceptable parts in the Tracker.
|
Although a bit of a perfectionist, I’m not disagreeing with you on this, but I
don’t wish to represent either view in the tuturial ... “pick it up as you go
along” ;-)
| |
Anyway Niels, as far as the actual text you wrote, the only problem I have
with it (besides POV-Ray & stud logos), is that a lot of stuff you wrote
regarding bowties, non-coplanar quads, & fixing BFC really apply to authors,
not to reviewers.
|
With a lot of parts, you’ll find that you have to do it yourself anyways because
of inactive authors. But I’m basically including it because the new reviewers
need to understand what they’re looking for, and because the explanation of
those words (scattered over several Ldraw pages) are really fuzzy:
I read those
back in the days, but still had to have the concepts explained to me in the PT
before I figured out what it was.
Now I’ve made that graphical representation I
really needed when I started parts authoring and generally roaming the PT.
It would be really nice to have this included in a parts authoring tutorial also
(or instead): but this is not the one, and until it’s here... well, I’m just
spreading a bit beyond the subject to patch up shortcomings elsewhere
| |
All the reviewers really need to know regarding l3p errors
is to compare errors found to this page:
http://www.ldraw.org/library/tracker/ref/l3pmsg/ and hold/warn accordingly.
The explanations you give would be good to have in a separate article such as
“What do those L3P -check errors really mean?”
|
I’ve already added link to that page
Anyways, thanks for the feedback, I think the tutorial will greatly benefit from
the general/overview comments you made.
Cheers NB
|
|
| |
In lugnet.cad, Niels Bugge wrote:
| |
Well, I don’t recall encouraging to place “illegal” hold votes in the
tutorial - yes, I write about BFC’ing and studlogos, but only as “best
practice” and associated with novotes...
|
Sorry about that, I didn’t mean to imply that you did. I just thought that list
might be useful for beginners.
|
|
| |
In lugnet.cad, Willy Tschager wrote:
> In lugnet.cad, Travis Cobbs wrote:
> > Just out of curiosity, have you tried using the latest ATI drivers? Or is it a
> > notebook that the manufacturer hasn't posted updated drivers for? (I know you
> > emailed me about this a long time ago, but can't remember the specifics.)
>
> the latter :-( the driver is two years old and was a one shot for those notebook
> series.
You may be able to install the Omega drivers on your laptop. They're tweaks of
the official ATI drivers, and claim to allow installation on all supported ATI
chipsets, not just the "Built by ATI" desktop chips. The Radeon 9x00 chips are
on the supported list. You can find them here:
http://www.omegadrivers.net/
I would strongly recommend reading the info on that site carefully before trying
the drivers, but I believe they will work.
--Travis
|
|
| |
In lugnet.cad, Willy Tschager wrote:
> In lugnet.cad, Niels Bugge wrote:
> > non coplanar quads ... can be particularly nasty to get rid of
>
> LDView checks for this errors and offers you a possible split into triangles. be
> aware: when you copy the error to the clipboard LDView will round the figures.
> as far as I know travis is working on this.
Non coplanar quads can also be caused by an error in one of the coordinates.
Please check the coordinates first before splitting the quad into triangles!
Niels Karsdorp
|
|
| |
In lugnet.cad, Niels Bugge wrote:
| |
In lugnet.cad, Andrew Westrate wrote:
| |
Anyway Niels, as far as the actual text you wrote, the only problem I have
with it (besides POV-Ray & stud logos), is that a lot of stuff you wrote
regarding bowties, non-coplanar quads, & fixing BFC really apply to authors,
not to reviewers.
|
With a lot of parts, you’ll find that you have to do it yourself anyways
because of inactive authors. But I’m basically including it because the new
reviewers need to understand what they’re looking for, and because the
explanation of those words (scattered over several Ldraw pages) are really
fuzzy:
I read those back in the days, but still had to have the concepts
explained to me in the PT before I figured out what it was.
Now I’ve made
that graphical representation I really needed when I started parts
authoring and generally roaming the PT.
|
If you (started as a reviewer on a part) start to do the edit of a part yourself
because of inactive authors, your function changes from part reviewer into
parts author. So how to fix those errors apply to authors, not to reviewers.
Reviewers should only know how to identify those errors.
Parts authors should know how to fix those errors.
About your reviewing tutorial, perhaps you can include a checklist, like
the one I wrote (I can’t remember if I mailed it to you or if I posted it
here).
Niels Karsdorp
|
|
| |
In lugnet.cad, Andrew Westrate wrote:
| |
Sorry about that, I didn’t mean to imply that you did. I just thought that
list might be useful for beginners.
|
np
|
|
| |
In lugnet.cad, Niels Karsdorp wrote:
| |
About your reviewing tutorial, perhaps you can include a checklist, like
the one I wrote (I can’t remember if I mailed it to you or if I posted it
here).
|
You did mail it to me: I’d really like to include it, but it seem a bit too
sketchy imo: I’ve mailed it back to you with some suggestions 2 times, but I
guess your spamfilter ate it or someting...
I’ll try to find it on my desktop when I turn it on next time
Cheers NB
|
|
|
