But, before, we should move it to a more publicly available repository. Perhaps ObjectProfile or Moose would be better. Doru On Thu, Mar 20, 2014 at 3:18 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote: -- www.tudorgirba.com "Every thing has its own flow"

On Mar 20, 2014, at 11:19, Tudor Girba <tudor(a)tudorgirba.com> wrote: Yep, that would be good. I think I will do it tomorrow. Right now I am working on delicate refactoring and restructuring on Graph-ET2. It involves creating higher level diagram builders for every chart (Lines, Histogram, Scatterplot ) and removing axis building knowledge from GET2DiagramAbstract hierarchy by putting it in a dedicated axis builder. Every diagram builder is in charge to have a default axis configuration considering its particularities. Then the GET2AbstractDiagram gets lighter and dedicated just to plotting the diagram content (bars, dots, lines, areas, etc). Cheers, Pablo

Great to hear that people are pushing GraphET Now I would really like that you try to exercise the library to really build graphics as we see in books. Stef On 20 Mar 2014, at 19:23, Pablo Estefó <pestefo(a)gmail.com> wrote:

Yes!

Hi, This is great. And it looks beautiful. Just, please make the labels gray by default :). I reviewed the code a bit: GET2Bar>>diagramClass ^ vertical ifTrue: [ GET2VerticalBarDiagram ] ifFalse: [ GET2HorizontalBarDiagram ] GET2Bar>>horizontal vertical := false I think you should change two things. First, you could probably work directly with the diagram instance and not with the diagram class. Second, instead of working with flags, just instantiate the correct object when you know it. For example: GET2Bar>>horizontal diagram := GET2HorizontalBarDiagram new Also, could you please rename GET2AbstractBuilder to GET2Builder? (abstract does not add much to it) In any case, I like it a lot that we have a RTBuilder hierarchy for everything related to visualizations. This should probably make it easy for creating fluent interfaces when wrapping in external interfaces like Glamour because the only thing we need to know from outside is to work with the interface from RTBuilder. One downside for the current implementation of the builders is that there is no distinction between the specification step and the rendering step (they are done in one shot). This makes things simpler now, but it has implications: - You cannot serialize the view specification object and render it in a different place (this for example, might be interesting for an Amber client and a Pharo server). - You cannot perform optimizations in the builder or work with things that are not yet defined. For example, in a Mondrian builder the order of defining nodes and edges should not necessarily matter. This should be possible: view edgesFrom: #superclass. view nodes: classes. This is only possible if we decouple the rendering from the specification. I think it would be worth it for us to spend a bit of time looking at this, but we can also probably do it at a later time. Doru On Wed, Apr 2, 2014 at 5:03 AM, Pablo Estefó <pestefo(a)gmail.com> wrote: -- www.tudorgirba.com "Every thing has its own flow"

Really nice. if this is the case I hope that there is a way to specify the labels color not to be grey :) I agree with the remarks below. On 02 Apr 2014, at 07:50, Tudor Girba &lt;tudor(a)tudorgirba.com&gt; wrote: > > I reviewed the code a bit: > GET2Bar>>diagramClass > ^ vertical > ifTrue: [ GET2VerticalBarDiagram ] > ifFalse: [ GET2HorizontalBarDiagram ] > > GET2Bar>>horizontal > vertical := false > > I think you should change two things. First, you could probably work directly with the diagram instance and not with the diagram class. > > Second, instead of working with flags, just instantiate the correct object when you know it. For example: > GET2Bar>>horizontal > diagram := GET2HorizontalBarDiagram new > > Also, could you please rename GET2AbstractBuilder to GET2Builder? (abstract does not add much to it) > > In any case, I like it a lot that we have a RTBuilder hierarchy for everything related to visualizations. This should probably make it easy for creating fluent interfaces when wrapping in external interfaces like Glamour because the only thing we need to know from outside is to work with the interface from RTBuilder. > > One downside for the current implementation of the builders is that there is no distinction between the specification step and the rendering step (they are done in one shot). This makes things simpler now, but it has implications: > - You cannot serialize the view specification object and render it in a different place (this for example, might be interesting for an Amber client and a Pharo server). > - You cannot perform optimizations in the builder or work with things that are not yet defined. For example, in a Mondrian builder the order of defining nodes and edges should not necessarily matter. This should be possible: > view edgesFrom: #superclass. > view nodes: classes. > This is only possible if we decouple the rendering from the specification. > > I think it would be worth it for us to spend a bit of time looking at this, but we can also probably do it at a later time. > > Doru > > > On Wed, Apr 2, 2014 at 5:03 AM, Pablo Estefó &lt;pestefo(a)gmail.com&gt; wrote: > Hi folks, > > I created some builders to ease chart configuration. You can use the previous builder as well but the idea is to have specific builders for specific charts. > Here you have some examples. > > <Screenshot 2014-04-01 18.29.31.png> > =.=.=.=.=.=.=.=.=.=.=.=.=. > |builder MAX| > MAX := 8. > builder := GET2Bar data: ((1 to: 15) collect: [ :v | MAX atRandom - 5 ]). > builder color: [ :val | |alpha | > alpha := [ :v | (v abs / MAX) ]. > val positive ifTrue: [ Color blue alpha: (alpha value: val) ] > ifFalse: [ Color purple alpha: (alpha value: val) ] > ]. > builder open > =.=.=.=.=.=.=.=.=.=.=.=.=. > > > <Screenshot 2014-04-01 18.03.29.png> > =.=.=.=.=.=.=.=.=.=.=.=.=. > | values builder | > values := #( #(1 3 3) #(3 2 4) #(5 14 5) #(6 10 1) #(7 9 6) #(8 8 9) ). > > builder := GET2Scatterplot data: values. > builder > color: (Color purple alpha: 0.5); > r: [:v | v third * 4]. > =.=.=.=.=.=.=.=.=.=.=.=.=. > > I am now fixing the Gofer configuration (for PDF exporting). > I’ll be working on better numbering strategy for axis and horizontal/vertical line guides as background. > > Stay tuned :-) > > Pablo > > > On Apr 1, 2014, at 3:27, Stéphane Ducasse &lt;stephane.ducasse(a)inria.fr&gt; wrote: > >> >>> Yes! >>> >>> I’ve started to use GraphET to generate figures in my papers. It works relatively well :-) >>> >>> But yes, we will continue to push it! >> >> Yes! >> >> >> _______________________________________________ >> Moose-dev mailing list >> Moose-dev(a)iam.unibe.ch >> https://www.iam.unibe.ch/mailman/listinfo/moose-dev > > > _______________________________________________ > Moose-dev mailing list > Moose-dev(a)iam.unibe.ch > https://www.iam.unibe.ch/mailman/listinfo/moose-dev > > > > > -- > www.tudorgirba.com > > "Every thing has its own flow" > _______________________________________________ > Moose-dev mailing list > Moose-dev(a)iam.unibe.ch > https://www.iam.unibe.ch/mailman/listinfo/moose-dev