Functionality board

General discussion about the development of the open source Blender

Moderators: jesterKing, stiv

dreamerv3
Posts: 119
Joined: Wed Oct 16, 2002 10:30 am

Post by dreamerv3 »

Yeah I'm gonna have to second that, in fact I'm starting to set up a website and post links, I'm not gonna torture these people with novels anymore, well only when absolutely neccessary... 8)

thorax
Posts: 320
Joined: Sun Oct 27, 2002 6:45 am
Contact:

Post by thorax »

joeri wrote:Thorax: I did not read any of your stuff.
Way to much text to be possibly interresting.

-Joeri
Note Joeri and I think IonPulse are coders.. Joeri is cool,
I like him.. I have nothing against him.. I don't know about IonPulse..
I'm not here to collect fans or enemies, just trying to loosen the crowd up and see if we can get some of the readers to respond, its something like one responder for every 10 readers, can't tell if they are unique reads or
just re-reads by the same users, but there is supposedly this
userbase of hundreds of thousands.. It makes me wonder why they
don't speak up.. What is with this crowd anyway..

Michel
Posts: 180
Joined: Wed Oct 16, 2002 7:27 pm
Location: Somewhere below the rivers in Holland (but not Limburg)

Post by Michel »

Hi,

I think something like the FuBo is a good idea. I personally would find it a great way to receive feedback on the things I do in the Blender sources. Such feedback will only improve the end-result in functionality and usability.
Doing a rewrite of a certain part of blender should not involve feedback from the FuBo in my opinion. Fellow developers provide feedback on parts that are rewritten. And trust me: they do :)

With regards,
Michel

beatabix
Posts: 72
Joined: Fri Oct 18, 2002 1:12 am

Post by beatabix »

IoN_PuLse wrote:
joeri wrote:Thorax: I did not read any of your stuff.
Way to much text to be possibly interresting.

-Joeri
Amen to that.
Yeah, thorax
don't you have a job or something to go to?
where do you find time to write all those four page replies?
try breaking it up into bite sized chunks, be direct and to the point.
you may get a few more responses than just those from the usual suspects.
reading is boring, this is the post-MTV era, we only have thirty second attention spans and we want everything five minutes ago.
later

BEAT

thorax
Posts: 320
Joined: Sun Oct 27, 2002 6:45 am
Contact:

Post by thorax »

Michel wrote:Hi,

I think something like the FuBo is a good idea. I personally would find it a great way to receive feedback on the things I do in the Blender sources. Such feedback will only improve the end-result in functionality and usability.
Doing a rewrite of a certain part of blender should not involve feedback from the FuBo in my opinion. Fellow developers provide feedback on parts that are rewritten. And trust me: they do :)

With regards,
Michel
Well rewrites I'm talking about are not really about rewriting what the
developers do, its more about rewriting what they are developing on,
the sources are sadly structured for the render/animation side of blender,
while the gameblender sources have had the benefit of good discussion in a commercial organization, its evidence of work of NaN.. To
make future changes there needs to be a pluggable architecture that
can scale, and blender is not 100% there yet..

thorax
Posts: 320
Joined: Sun Oct 27, 2002 6:45 am
Contact:

Post by thorax »

beatabix wrote:
IoN_PuLse wrote:
joeri wrote:Thorax: I did not read any of your stuff.
Way to much text to be possibly interresting.

-Joeri
Amen to that.
Yeah, thorax
don't you have a job or something to go to?
where do you find time to write all those four page replies?
try breaking it up into bite sized chunks, be direct and to the point.
you may get a few more responses than just those from the usual suspects.
reading is boring, this is the post-MTV era, we only have thirty second attention spans and we want everything five minutes ago.
later

BEAT
Its also the age of short-term relationships and poor sense of identity,
I don't see short attention spans as an excuse..

Michel
Posts: 180
Joined: Wed Oct 16, 2002 7:27 pm
Location: Somewhere below the rivers in Holland (but not Limburg)

Post by Michel »

thorax wrote:
Michel wrote: Doing a rewrite of a certain part of blender should not involve feedback from the FuBo in my opinion. Fellow developers provide feedback on parts that are rewritten. And trust me: they do :)
Well rewrites I'm talking about are not really about rewriting what the
developers do, its more about rewriting what they are developing on,
the sources are sadly structured for the render/animation side of blender,
while the gameblender sources have had the benefit of good discussion in a commercial organization, its evidence of work of NaN.. To
make future changes there needs to be a pluggable architecture that
can scale, and blender is not 100% there yet..
Hi,

Yes, I understand that you're talking about a redesign of blender. I tried to stay away from that in my post, but I'd better explain I guess.
What Ton is talking about is a functionality board for reviewing new features that are developed. Such a review can (and needs to) be done at two levels:
1 - user level. Is the new functionality indeed useful. Are there any small modifications needed for it to be even more useful.
2 - source level. Is the design and code fitting in the current source tree.

The 1st point should be done by users, and the 2nd point by fellow developers.
What you're talking about is a complete redesign of blender which is not on this topic. Please stay to the topic started by Ton and continue the redesign discussions in the countless other topics (started by you).

With regards,
Michel

IoN_PuLse
Posts: 212
Joined: Wed Oct 16, 2002 6:05 am
Location: Canada, BC
Contact:

Post by IoN_PuLse »

FYI:

I am a coder, and have programmed in C for some time and learned C++ completely about a year ago (and have been using it since)

The reason I'm not bothering to read the huge monologues is because a) I don't have the time and b) I'm sure it could be expressed in a more compact manner. But hey, that's just me.

thorax
Posts: 320
Joined: Sun Oct 27, 2002 6:45 am
Contact:

Post by thorax »

ArgoUML seems to be okay, works just like Rational Rose.. This is valuable experience if you plan to work in a major outfit someday
and are required to use CASE tools in you rdevelopment..
So get ArgoUML, this is what the designers, designer-coders and coders will use to communicate ideas.. The users can use ArguUML if they
want, maybe to work on USeCase's, but its really for designers and coders.

I did a test use-case of Verse in ArgoUML, saved out and loaded it up fine.. It seems to work, just needs a lot of ram and a good processor..
Its pure java..

I am going to read through the blender source and glean what I can
of the modules and functions, and try to produce objects of this and define them as classes in Argo.. That should help in giving a general
overview of the source code.. There was a diagram called a collaboration diagram that I'm not familiar with..


What I ask of you, users to look at blenders interface and determine what are inconsistencies in the interface, like if a feature exists somewhere document where it should also exist in the interface.. Also look to find the features that are alike and features that should be alike. This is a method of determining how things are related in the interface or should be and how to determine the code that will need to be turned into objects when blender goes to C++. An example is like say you want the magnet tool to work in Lattices as it works on meshes, you would say that..
This helps us determine features that are related or should be related..


Now for the coders, the idea there is to determine what some of the code
does, if you can comment what your understand, what you don't don't..
If you want to make suggestions on the sources, like design oriented
suggestions (things you would like to change but are outside the scope
of what you are focusing on) you might want to write up a document on
it, or make some UML graphs.. The UML and docs are how to express
ideas that affect the entire source.. If you see some source that is
really screwy and causing you headaches for whatever reason, document it and specify where it occurs.. It won't be possible to use the UML to right away to show the problems in a graph..

IFIf anyone knows of any programs that take source and generate relationship diagrams, it would be good to speak up.. This could
be used to highlight modules that are causing problems in the source tree, if for yourself or to communicate to others, without a big wordy discussion.

I think this would be a good start on the redesign/refactor of the blender sources..

Eventually I see that some kind fo web-interactive daigraming and communication system could be created to help the developers
focus attention on daily issues.. Like being able to put big red X's
on objects that are conflicting with your source.. Its better to do that
than go into the other guys sources changing his stuff out from under him..

matt_e
Posts: 410
Joined: Mon Oct 14, 2002 4:32 am
Location: Sydney, Australia
Contact:

Post by matt_e »

Thorax: Please keep your posts on topic.

What you just wrote has very little relevance to Ton's asking for feedback for a proposed functionality board. If you can't stay on topic, I'm going to have to start deleting posts and posting them in separate threads, which I don't want to do, but these threads are veering and rambling all over the place making it confusing and incoherent for everyone who is trying to read them.

p_vertex
Posts: 0
Joined: Mon Jan 27, 2003 1:45 am

Post by p_vertex »

I generally think this is a good idea, if for no other reason that to facilitate some kind of communication between coders and users. I've read all the posts here, but I'd like to take issue with B@rt's notion that the 'coders are here to do what they want, and if you want it done better, do it yourself.' I'd like to give the programmers more credit - I'd like to think that their objective is to contribute in a thoughtful and positive way, such that the result is something useful. In order to be useful, it requires finding out what people need, and what kinds of tools, usability issues, etc, will make the overall effort a successful one.

Overall, the functionality board will add some balance to the process so that it's not biased toward the mindset of the coder. This is not intended in any way as a sign of disrespect - I know from my own experience how easy it is to create a mental image of what someone has in mind (typically a user), only to find that my interpretation was a bit off the mark. Hopefully, a functionality board will help to minimize the potential for this to occur.

thorax
Posts: 320
Joined: Sun Oct 27, 2002 6:45 am
Contact:

Post by thorax »

Well the problem I have as a coder is determining what is the best way, its too easy to make half-assed design decisions in the middle of the code when just doing whatever I think I need (called hacking).. It is possible to do your own designs and code against those, but blender's source is just too large for a single coder to be able to make mass decisions about the source code, this requires groups of developers to make decisions,
and usually it involves looking at a design, either with pseudocode or
graphs or both..

But its not going to be possible for people who are wishing stuff into the
interface to influence the coding design without getting somewhat dirty, but it doesn't require coding.. It requires structuring thought about the
features.. The users must at least suggest features in some way that
can be categorized easily.

I'm finishing off some PHP code that I hope will help out here.
The users will be able to point at features that exist in blender,
draw boxes around stuff in the interface and comment on what they
think should exist there.. I can then categorize the feature suggestions
by the previous feature reference.. Then I and other designers could collect the features together and determine what are the shared features, and determine a structure to support that feature.. Overtime we could collect similarly defined features that support similar structures..

( Note, however that what I have so far in this annotation program
is suitable for only pointing at interface elements and referencing
the features with URL's.. You'll see what I mean when I release it..
Its just a idea I had.. )

This, the design docs and graphs and categorization of features, will help the coders to make informed decisions about what structures can support the features users generally request (or to support features they would like themselves), but it has to be presented to the coders in a way that is easy to access..There will be no way the developers are going to want to
talk to the users because the users have no concept of what is easy or
hard, but the coders will like the design docs because it gives
them some perspective on what needs to be done without talking to hundreds of users.

(I'm trying to cut down this message.. But its hard to say everything
I'm thinking simply)..

cessen
Posts: 109
Joined: Tue Oct 15, 2002 11:43 pm

Post by cessen »

Sorry to be off-topic for a moment, but I'd like to point out that being concise is actually not very easy, and it takes a lot of practice and effort to get good at it. I am a prime example of someone who needs such practice. :-)
Of course, I'm not as bad as Karl Marx, who takes about ten pages to say things that I can say in a short paragraph. ;-)

As far as Thorax being off-topic goes, I'm not so sure that he is. Granted, his posts are round-and-about and confusing, but I think what he's driving at is a means for coders and designers/users to communicate with eachother (this "ArgoUML" thing). That certainly seems related to the original post.

It would be nice if he were clearer and more to-the-point, but as I said, that is something that comes with practice (just like all communication skills).

thorax
Posts: 320
Joined: Sun Oct 27, 2002 6:45 am
Contact:

Post by thorax »

cessen wrote: As far as Thorax being off-topic goes, I'm not so sure that he is. Granted, his posts are round-and-about and confusing, but I think what he's driving at is a means for coders and designers/users to communicate with eachother (this "ArgoUML" thing). That certainly seems related to the original post.

It would be nice if he were clearer and more to-the-point, but as I said, that is something that comes with practice (just like all communication skills).
Thanks!!

UML stands for Unified Modelling Language, when something is
unified it is "made into a unit".

An example of UML is shown below
Image

Now notice how easy this is to observe as compared with trying to
read source code, the example shown is read:

A polygon object class inherits (the triangle symbols mean inherit from, the arrow is in the direction of what is inherited) from two classes "OneDimensional" and "TwoDimensional".. These two object classes inherit from the "Shape" class..
Oh, And the polygon class includes (or uses, or is associated with) a set of (1..*, or at least one to many) of vertices defined with x,y integers..

There is more information in this diagram that I can describe but I
don't want to bore you guys..

dittohead
Posts: 122
Joined: Mon Oct 14, 2002 2:49 am

Post by dittohead »

ton wrote:Hrms. Confusing feedback and with mixed results... i would prefer if posters stay on topic, and be as brief as possible in replies.

My only aim was to find a group of users (coders, artists) who like to help out in spending time on specs, feature requests and improvements.

I'll put the idea back on the drawing board.
this project essentially make it easy for non-coders to get their ideas on the table.

i like the idea.

my ideas:
  • #- Weekly irc meet(s)(with coders sometimes). this will allow us to put idea forwad for the board here in the beginning, and in later stages allow us to put idea forward for blender.
    #- (larger)ideas that most(or all)of us agree on could be sent to a poll, and a news item at elysiun, and a new item here on blender.org for general consumption by the blender user masses.
    #- a site where one could upload designs/ideas that require their own page, like cessen's material overhaul, could be published for us--or others--to see, and comment on. we could improve each oher ideas this way as well.
i'll add more to this when i get more ideas.

i also think we should stay on topic as ton asked us to. :roll:
dittohead

Post Reply