Day after day, picture after picture, following the weather and seasons
More than 300 "Pictures of the day" on, my personal page

Latest articles
2013, June 30Auto-generated 3D Models - Appendix B1 - A few experiments with a 3D Machine
2013, June 18Auto-generated 3D Models - Appendix B - Meet The BIG-3D-GENERATOR Machine
2013, June 6Auto-generated 3D Models - Appendix C - Our Old Models Now Very Slim
2013, May 25Auto-generated 3D Models - Appendix A - About St. John auto-generated
2013, May 14Auto-generated 3D Models - 3) Cezanne's Lesson
2013, May 4Auto-generated 3D Models - 2) Visual Comparison
2013, April 29Auto-generated 3D Models - 1) Environment
2013, April 10Thanks cgilbert!
where it is shown how, in certain cases, an evaluation with four stars counts much more then... five stars!
2013, April 7Hi Joe (Joes)...
3D Warehouse fake users customs and traditions : some things that not everyone knows...
2013, March 29StPeter on "El País"
Marc 12 last the Spanish newspaper "El País" published my 3D model of St Peter's Basilica with the Bernini's Colonnade - but they did not tell me not even "thanks"!

Wednesday, June 16, 2010


Classifying 3D warehouse Evaluations
Enrico Dalbosco (Arrigo Silva) - Padua, June 2010

What do the stars say
to Don Quixote during
his Vigil Of Arms?

What do they say
to All The People
throughout infinite nocturnal sky??

What did the stars say
to Poets Wizards
Navigators Astrologers
Scientists Astronomers
Night Owls
Nightingales Drunkards???

Who knows…

Often, alas, they stay quietly silent.


I tried to make this classification in a strictly objective, without questioning the sincerity of assessments or the merits of the evaluator intentions or psychology etc. but only basing on the analysis of words and phrases actually present in 3D warehouse evaluations.
Some other topics that might be explored: relationship between evaluations and so called 3D warehouse rating; classification of users based on their evaluations; cluster analysis on user cross-evaluations.


The following table illustrates the main categories of evaluations present in 3D Warehouse. On each row I have brought a separate category, identified by a code (which I will use later) and characterized on the number of stars (5, 1 to 4, or 0 stars) and the type of evaluation.

CategoryItemRatingReview Typical Content - An Example
 A-Silentno words
not specific [too generic / too laudatory / others]
Great! / The best of the best! / Vote my models! /See you next week...
 A-Specific(pos)specific > positive
I like these excellent textures because...
 A-Specific(mix)specific > positive with some criticisms 
Very well structured and light, even if there are...
B-Silentno words
 B-Genericnot specific
I dislike your model...
 B-Grimacespecific > positive (?) or negative with some appreciations (?)
Nice textures... - Bad model with nice textures...
 B-Specific(neg)specific > negative
Poor textures and misaligned...
C-Silent  (no stars)no words
 C-Generic  (no stars)not specific
Visit my site at... / See you next month...
 C-Specific(pos)  (no stars)specific > positive
I like these excellent textures because...
 C-Specific(mix)  (no stars)specific > not positive
Poor textures...

Let me now express some thoughts that I have gained about the various types of assessment.


The more frequent evaluations are those of category A-Silent and A-Generic that do not bring any contribution to a critical analysis of the contents of the model. The evaluations A-Gdeneric often denote a marked laudatory attitude without helping us to understand on wich strength of the model it is based (I would call this sub-category as... G-Wow!).

Less frequent but much more interesting are A-Specific (positive) evaluations, and even less frequent A-Specific (mixed) evaluations, that  instead help the model author (and all the readers, thus including all 3D warehouse modelers) to understand what are the aspects considered more (or less) valid by the evaluator.


I have classified as type B all evaluations form 1 to 4 stars involving a partially or completely non positive assessment of the model by the assessor.

Much frequent are the B-Silent and B-Generic evaluations that do not bring any contribution to a critical analysis of the contents of the model, unlike those B-Specific (negative) that can help us understand why the assessment is not positive.

A special consideration deserve B-Grimace evaluations because they contain a contradiction in terms... Here I do not want to explain my personal judgement, and leave to the acuity and the perspicacity of the readers draw their own conclusions.


This types of evaluations are relatively infrequent. In some cases they depend on evaluator mistake, in othe they are messages (less or more undestandable) that the evaluator wants to send to model Author or to the 3D warehouse colleagues…


1 comment:

alsomar said...

A little comment about null evaluations: I think you can't give a 'C-Silent' rating, because a warning text appears saying: 'Please enter a rating or review in order to save'.

Anyway, many thanks for share your thoughts and tutorials!