Subject:
|
Re: I want to become a reviewer
|
Newsgroups:
|
lugnet.cad
|
Date:
|
Tue, 5 Oct 2004 12:34:11 GMT
|
Highlighted:
|
(details)
|
Viewed:
|
1034 times
|
| |
|
|
Hi Niels
I think it’s really, really great that you whish to join the core. (hey guys,
give him a hero’s welcome!)
I myself haven’t been a reviewer for very long (or haven’t authored many parts
yet for that sake) and isn’t particularly experienced, but I don’t see this as a
hindrance to giving advice, as this only means that I have a lot of rookie
experiences in fresh memory.
First of all relax! Reviewing is nothing different than the error checking you
usually do before submitting a part; it’s all about getting enough eyes on the
part to find all errors. The only trouble is you have to have somebody else (the
author) fix the errors you find. You could consider refraining from casting hold
votes (and certification votes) at first, to avoid making errors and upsetting
people until you get more experience: I personally haven’t cast any hold votes
yet (except on one of my own parts), but have to begin casting certification
votes quickly (see why below), but remember: Two other reviewers has to cast one
too before the file gets released, so you can’t screw up badly (and your vote
gets deleted if the file is uploaded again because somebody else has detected an
error, and lastly the admins need to accept it as well, and then it’s their
responsibility!)
I usually run four checks on the file: The first one is running the file through
in mlcad with “draw to selection only” turned on (the button next to the
gridbuttons) to get a sense of what objects the file is made up of; it’s here I
find missing things (like edges), extra things that need to be deleted (usually
objects made obsolete by more advanced primitives) and highlighting (changing
the colour of) things that need to be changed via other programs like: L3lab
where I look for BFC-errors (turning “mytest6” on in the testmenu), and LDDP
where I fix the windings and run l3p-check to find bad vertices (you have to
turn non coplanar detection on manually in the tools/options/configuration
values-menu – I can’t figure out why it isn’t turned on by default..). Finally I
render the part in Pov-ray (via L3DAO) to check the part for gaps (where planes
don’t meet creating a black line) and if the studlogos is turning the right way.
Note to parts authors: Parts without a logical sequence of objects (like an
onion with inner parts first and outer ones last so nothing gets in the way of
other) is a pain in the .. as they’re impossible to figure out, and DON’T group
all the lines in one section of the part making a mess of lines, but keep them
together with the planes they frame; then it’s possible to see what belongs to
each other and thereby what’s extra or missing. Cf.
x250 which is very
nice although it could do with some rect-substitution ;-)
That was something about checklists and tools, but I’ve also got some
experiences from the PT about some important do’s and don’t’s I’d like to share:
#1 Don’t touch!
It’s the authors part and if you put your dirty hands on it, it creates an extra
0 line with your name on it and blocks your ability to certify it, meaning
that with one out, 4 reviewers have to look at it before it can be released,
and that can mean a lot of extra time in the PT (there’s too few reviewers
already). And the author gets pissed, so don’t touch unless it’s absolutely
nessecary.
#2 Always follow up:
Except for a reviewer hijacking your part, there’s nothing worse for an author
that a reviewer that just leaves a comment about an error, but doesn’t return
to certify (or pointing at other errors) when you upload a new fixed version
of the file: He has used a lot of time making it, and now it’s even better,
but nowhere closer to actually getting released: It’s very bad for the morale
of the author (why should he react to reviews or even make new parts?), and
the part sinks down to the dusty graveyard of neglected parts.
I’m sure that it isn’t out of bad will that the reviewers don’t return, but only
because they respond to too many parts, so it’s crucial to:
#3 Limit your activity to a few parts:
So you can follow up on them, here the automatic notices from the PT a really
important tool, where the message of a new upload only appears once, and
should therefore prompt another visit and update of your vote: If you don’t,
the part goes to the graveyard, because you forget the information you
achieved about the part and have to “re-learn” it if you return at a much
later date. An example of this is a (granted: difficult) part
6253s-01.dat I’ve
witnessed:
8 reviewers has been in over the part, each using time on reviewing the part
(some more than once), I’ve used a day or two to fix it (sbliss has worked on
it too), but as everybody has forgotten it and failed to return, most of the
reviewers work is wasted as three (new) reviewers has to come in and
“re-learn” the part in order to be able to certify it: It could be certified
by now or even released with all those reviewers involved.
To sum it up, my opinion is that the ideal way for authors and reviewers to
cooperate is to work the part as a fast ping-pong between them and follow the
part all the way to the door while it’s in fresh memory of both. If one of them
leaves the game, a lot of knowledge and work is lost, and the longer the part
stays on the cemetery, the bigger the risk is that the author looses interest
and abandon the part or authoring altogether.
Furthermore have a suggestion that groups of reviewers team up to work with a
part, as a kind of “fast track” reviewing to keep the author “warmed up”. Is
there 3-4 volunteers that want to try it out? And by the way; how many reviewers
are there actually?
Cheers
The other Niels
|
|
Message has 3 Replies:
 | | Re: I want to become a reviewer
|
| (snip) Wow, that's some great advice. Lots of good things for new reviewers to take heed and consider. I'd love to see that info extracted and made referencable, perhaps somewhere in this area: (URL) specifically either here (URL) or here (URL) (...) (20 years ago, 5-Oct-04, to lugnet.cad, FTX)
| | | Re: I want to become a reviewer
|
| Excellent post, Niels! :) (...) To expand on this- it's very good to keep edges grouped with polygons, but it's also nice (at least at the beginning of the file), to put the edges first. Laying out the frame of the part makes it easier to see what (...) (20 years ago, 5-Oct-04, to lugnet.cad, FTX)
| | | Re: I want to become a reviewer
|
| (...) How imported is the orrientation of the studlogos? Should these face exactly the same direction as the real part? Or do they just have to be readable (thus not mirrored) and all studlogos on one part face the same direction (unless the real (...) (20 years ago, 3-Nov-04, to lugnet.cad, FTX)
|
Message is in Reply To:
| | I want to become a reviewer
|
| Over time I have increased my experience in parts authoring and still learning new things. But with all new and improved parts added to the PT the list of parts that need to be reviewed is ever increasing. So I'm thinking to become a reviewer, too. (...) (20 years ago, 2-Oct-04, to lugnet.cad)
|
25 Messages in This Thread:
  
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
