[svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
One non-requirement: I don't care if the vertices partially cover the edges. The edges will be painted first, and will run from vertex to vertex as straight as the crow flys. Then the vertices will be painted on top of them. Any hints or guidance appreciated. Thanks. --- In svg-developers@yahoogroups.com, "run2bmi21" <[EMAIL PROTECTED]> wrote: > > Hi, > > I was wondering whether anyone knows of a good source for mapping > algorithms. > > I'm trying to map graph vertices (which in my specific case > represent both client and server computers, each with a specific IP > address) onto a 2-D SVG canvas. The vertices are represented by > circles of varying sizes. Lines are used to depict graph edges > (which in this case represent client-server connections). > > A nice algorithm would place the vertices onto the SVG canvas > (loosely by IP address) so that: > > 1. They don't overlap > 2. They're not too crowded, i.e. they've got a bit of elbow-room > 3. They're not too widely separated (even if their IP address *are* > far apart, i.e. there's not huge amounts of white-space on the > canvas) > 4. Vertices with similar addresses are positioned close to each other > 5. Vertices with dissimilar addresses are not as close together > > I realize that these requirements can't be met 100%. I guess I'm > looking for an algorithm that might be a good compromise. > > One other feature of the graph is that it will be nearly bipartite, > a subgraph of clients with no connections between the clients, and a > subgraph of servers with a very small number of connections between > the servers, but plenty of connections between the two subgraphs. > > I'm thinking a nice arrangement might be to position "server" > vertices in the center of the SVG canvas, and "client" vertices > surrounding them. > > If you can give me any tips, thanks. > > Doug > - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
[svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
Take a look at the following foaf example from Jim Ley. http://jibbering.com/foaf/foafnaut.svg Are you doing avoidance client side or static generation side? Static collision avoidance is a much more complicated issue, especially if you don't have the use of a nice SVG DOM. Alastair --- In svg-developers@yahoogroups.com, "run2bmi21" <[EMAIL PROTECTED]> wrote: > > One non-requirement: I don't care if the vertices partially cover > the edges. The edges will be painted first, and will run from > vertex to vertex as straight as the crow flys. Then the vertices > will be painted on top of them. > > Any hints or guidance appreciated. Thanks. > > --- In svg-developers@yahoogroups.com, "run2bmi21" > <[EMAIL PROTECTED]> wrote: > > > > Hi, > > > > I was wondering whether anyone knows of a good source for mapping > > algorithms. > > > > I'm trying to map graph vertices (which in my specific case > > represent both client and server computers, each with a specific > IP > > address) onto a 2-D SVG canvas. The vertices are represented by > > circles of varying sizes. Lines are used to depict graph edges > > (which in this case represent client-server connections). > > > > A nice algorithm would place the vertices onto the SVG canvas > > (loosely by IP address) so that: > > > > 1. They don't overlap > > 2. They're not too crowded, i.e. they've got a bit of elbow-room > > 3. They're not too widely separated (even if their IP address > *are* > > far apart, i.e. there's not huge amounts of white-space on the > > canvas) > > 4. Vertices with similar addresses are positioned close to each > other > > 5. Vertices with dissimilar addresses are not as close together > > > > I realize that these requirements can't be met 100%. I guess I'm > > looking for an algorithm that might be a good compromise. > > > > One other feature of the graph is that it will be nearly bipartite, > > a subgraph of clients with no connections between the clients, and > a > > subgraph of servers with a very small number of connections > between > > the servers, but plenty of connections between the two subgraphs. > > > > I'm thinking a nice arrangement might be to position "server" > > vertices in the center of the SVG canvas, and "client" vertices > > surrounding them. > > > > If you can give me any tips, thanks. > > > > Doug > > > - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
[svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
Thanks Alastair for the very interesting link. The kind of mapping I'd be interested in though is "static generation side" I think. I'm not interested in dragging objects or creating new vertices or edges dynamically on the SVG canvas. I'm interested in creating a huge static picture (a graph) of a client-server universe that contains about 1,500 vertices and about 5,000 edges, with none of the vertices hiding or overlapping each other (hidden or overlapped edges I don't care about), and with the vertices positioned on the 2-D SVG canvas roughly approximate to their position in the linear space of IP addresses. Linear IP address space vs. 2-D SVG canvas, minimal white-space, no vertex overlap: that's the problem. If I wanted to do a Google search to explore and maybe find mapping algorithms of this sort, what keywords would be good? Are you familiar with this area? Thanks again for responding, Doug BTW, Check out visualthesaurus.com for a really nice implementation of an interactive graph. The graph at http://jibbering.com/foaf/foafnaut.svg reminded me of it. --- In svg-developers@yahoogroups.com, "Alastair Fettes" <[EMAIL PROTECTED]> wrote: > > Take a look at the following foaf example from Jim Ley. > > http://jibbering.com/foaf/foafnaut.svg > > Are you doing avoidance client side or static generation side? Static > collision avoidance is a much more complicated issue, especially if > you don't have the use of a nice SVG DOM. > > Alastair > > > --- In svg-developers@yahoogroups.com, "run2bmi21" > <[EMAIL PROTECTED]> wrote: > > > > One non-requirement: I don't care if the vertices partially cover > > the edges. The edges will be painted first, and will run from > > vertex to vertex as straight as the crow flys. Then the vertices > > will be painted on top of them. > > > > Any hints or guidance appreciated. Thanks. > > > > --- In svg-developers@yahoogroups.com, "run2bmi21" > > <[EMAIL PROTECTED]> wrote: > > > > > > Hi, > > > > > > I was wondering whether anyone knows of a good source for mapping > > > algorithms. > > > > > > I'm trying to map graph vertices (which in my specific case > > > represent both client and server computers, each with a specific > > IP > > > address) onto a 2-D SVG canvas. The vertices are represented by > > > circles of varying sizes. Lines are used to depict graph edges > > > (which in this case represent client-server connections). > > > > > > A nice algorithm would place the vertices onto the SVG canvas > > > (loosely by IP address) so that: > > > > > > 1. They don't overlap > > > 2. They're not too crowded, i.e. they've got a bit of elbow-room > > > 3. They're not too widely separated (even if their IP address > > *are* > > > far apart, i.e. there's not huge amounts of white-space on the > > > canvas) > > > 4. Vertices with similar addresses are positioned close to each > > other > > > 5. Vertices with dissimilar addresses are not as close together > > > > > > I realize that these requirements can't be met 100%. I guess I'm > > > looking for an algorithm that might be a good compromise. > > > > > > One other feature of the graph is that it will be nearly bipartite, > > > a subgraph of clients with no connections between the clients, and > > a > > > subgraph of servers with a very small number of connections > > between > > > the servers, but plenty of connections between the two subgraphs. > > > > > > I'm thinking a nice arrangement might be to position "server" > > > vertices in the center of the SVG canvas, and "client" vertices > > > surrounding them. > > > > > > If you can give me any tips, thanks. > > > > > > Doug > > > > > > - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
[svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
It sounds to me like it's close to the labeling problem. Hava a look at the Java Topology Suit (http://www.vividsolutions.com/jts/JTSHome.htm) and see if that can give any directions. It's open source and pretty well documented. /Roger --- In svg-developers@yahoogroups.com, "run2bmi21" <[EMAIL PROTECTED]> wrote: > > Thanks Alastair for the very interesting link. > > The kind of mapping I'd be interested in though is "static generation > side" I think. I'm not interested in dragging objects or creating new > vertices or edges dynamically on the SVG canvas. I'm interested in > creating a huge static picture (a graph) of a client-server universe > that contains about 1,500 vertices and about 5,000 edges, with none of > the vertices hiding or overlapping each other (hidden or overlapped > edges I don't care about), and with the vertices positioned on the 2-D > SVG canvas roughly approximate to their position in the linear space > of IP addresses. Linear IP address space vs. 2-D SVG canvas, minimal > white-space, no vertex overlap: that's the problem. > > If I wanted to do a Google search to explore and maybe find mapping > algorithms of this sort, what keywords would be good? Are you > familiar with this area? > > Thanks again for responding, > Doug > > BTW, Check out visualthesaurus.com for a really nice implementation of > an interactive graph. The graph at > http://jibbering.com/foaf/foafnaut.svg reminded me of it. > > > --- In svg-developers@yahoogroups.com, "Alastair Fettes" > <[EMAIL PROTECTED]> wrote: > > > > Take a look at the following foaf example from Jim Ley. > > > > http://jibbering.com/foaf/foafnaut.svg > > > > Are you doing avoidance client side or static generation side? Static > > collision avoidance is a much more complicated issue, especially if > > you don't have the use of a nice SVG DOM. > > > > Alastair > > > > > > --- In svg-developers@yahoogroups.com, "run2bmi21" > > <[EMAIL PROTECTED]> wrote: > > > > > > One non-requirement: I don't care if the vertices partially cover > > > the edges. The edges will be painted first, and will run from > > > vertex to vertex as straight as the crow flys. Then the vertices > > > will be painted on top of them. > > > > > > Any hints or guidance appreciated. Thanks. > > > > > > --- In svg-developers@yahoogroups.com, "run2bmi21" > > > <[EMAIL PROTECTED]> wrote: > > > > > > > > Hi, > > > > > > > > I was wondering whether anyone knows of a good source for mapping > > > > algorithms. > > > > > > > > I'm trying to map graph vertices (which in my specific case > > > > represent both client and server computers, each with a specific > > > IP > > > > address) onto a 2-D SVG canvas. The vertices are represented by > > > > circles of varying sizes. Lines are used to depict graph edges > > > > (which in this case represent client-server connections). > > > > > > > > A nice algorithm would place the vertices onto the SVG canvas > > > > (loosely by IP address) so that: > > > > > > > > 1. They don't overlap > > > > 2. They're not too crowded, i.e. they've got a bit of elbow-room > > > > 3. They're not too widely separated (even if their IP address > > > *are* > > > > far apart, i.e. there's not huge amounts of white-space on the > > > > canvas) > > > > 4. Vertices with similar addresses are positioned close to each > > > other > > > > 5. Vertices with dissimilar addresses are not as close together > > > > > > > > I realize that these requirements can't be met 100%. I guess I'm > > > > looking for an algorithm that might be a good compromise. > > > > > > > > One other feature of the graph is that it will be nearly bipartite, > > > > a subgraph of clients with no connections between the clients, and > > > a > > > > subgraph of servers with a very small number of connections > > > between > > > > the servers, but plenty of connections between the two subgraphs. > > > > > > > > I'm thinking a nice arrangement might be to position "server" > > > > vertices in the center of the SVG canvas, and "client" vertices > > > > surrounding them. > > > > > > > > If you can give me any tips, thanks. > > > > > > > > Doug > > > > > > > > > > - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
[svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
--- In svg-developers@yahoogroups.com, "run2bmi21" <[EMAIL PROTECTED]> wrote: > > I'm interested in > creating a huge static picture (a graph) of a client-server universe > that contains about 1,500 vertices and about 5,000 edges, with none of > the vertices hiding or overlapping each other The most common tool for creating non-overlapping hierarchical directed graphs is graphviz, which can create SVG output, as well as graphfiles: http://www.graphviz.org/ Graphviz is in C; depending on your server software language, there is also a Java binding at: http://www.loria.fr/~szathmar/off/projects/java/GraphVizAPI/index.php I think you could use graphviz's concept of subgraphs to group your addresses by network. I'm not sure about guaranteeing strict ordering, though. Thanks, --kirby - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
Re: [svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
Hi, Did you look into these sites? http://directory.google.com/Top/Science/Math/Combinatorics/Software/Graph_Drawing/ i think this one sounds good: http://aragorn.ads.tuwien.ac.at/AGD/ please keep us informed, i'm also looking into this sort of thing. cya. Omar run2bmi21 schrieb: > Thanks Alastair for the very interesting link. > > The kind of mapping I'd be interested in though is "static generation > side" I think. I'm not interested in dragging objects or creating new > vertices or edges dynamically on the SVG canvas. I'm interested in > creating a huge static picture (a graph) of a client-server universe > that contains about 1,500 vertices and about 5,000 edges, with none of > the vertices hiding or overlapping each other (hidden or overlapped > edges I don't care about), and with the vertices positioned on the 2-D > SVG canvas roughly approximate to their position in the linear space > of IP addresses. Linear IP address space vs. 2-D SVG canvas, minimal > white-space, no vertex overlap: that's the problem. > > If I wanted to do a Google search to explore and maybe find mapping > algorithms of this sort, what keywords would be good? Are you > familiar with this area? > > Thanks again for responding, > Doug > > BTW, Check out visualthesaurus.com for a really nice implementation of > an interactive graph. The graph at > http://jibbering.com/foaf/foafnaut.svg reminded me of it. > > > --- In svg-developers@yahoogroups.com, "Alastair Fettes" > <[EMAIL PROTECTED]> wrote: >> Take a look at the following foaf example from Jim Ley. >> >> http://jibbering.com/foaf/foafnaut.svg >> >> Are you doing avoidance client side or static generation side? Static >> collision avoidance is a much more complicated issue, especially if >> you don't have the use of a nice SVG DOM. >> >> Alastair >> >> >> --- In svg-developers@yahoogroups.com, "run2bmi21" >> <[EMAIL PROTECTED]> wrote: >>> One non-requirement: I don't care if the vertices partially cover >>> the edges. The edges will be painted first, and will run from >>> vertex to vertex as straight as the crow flys. Then the vertices >>> will be painted on top of them. >>> >>> Any hints or guidance appreciated. Thanks. >>> >>> --- In svg-developers@yahoogroups.com, "run2bmi21" >>> <[EMAIL PROTECTED]> wrote: Hi, I was wondering whether anyone knows of a good source for mapping algorithms. I'm trying to map graph vertices (which in my specific case represent both client and server computers, each with a specific >>> IP address) onto a 2-D SVG canvas. The vertices are represented by circles of varying sizes. Lines are used to depict graph edges (which in this case represent client-server connections). A nice algorithm would place the vertices onto the SVG canvas (loosely by IP address) so that: 1. They don't overlap 2. They're not too crowded, i.e. they've got a bit of elbow-room 3. They're not too widely separated (even if their IP address >>> *are* far apart, i.e. there's not huge amounts of white-space on the canvas) 4. Vertices with similar addresses are positioned close to each >>> other 5. Vertices with dissimilar addresses are not as close together I realize that these requirements can't be met 100%. I guess I'm looking for an algorithm that might be a good compromise. One other feature of the graph is that it will be nearly bipartite, a subgraph of clients with no connections between the clients, and >>> a subgraph of servers with a very small number of connections >>> between the servers, but plenty of connections between the two subgraphs. I'm thinking a nice arrangement might be to position "server" vertices in the center of the SVG canvas, and "client" vertices surrounding them. If you can give me any tips, thanks. Doug > > > > > > - > To unsubscribe send a message to: [EMAIL PROTECTED] > -or- > visit http://groups.yahoo.com/group/svg-developers and click "edit my > membership" > > Yahoo! Groups Links > > > > > > > - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
Re: [svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
And if it helps: http://www.cs.brown.edu/people/rt/papers/gd-tutorial/gd-constraints.pdf Omar Abo-Namous schrieb: > Hi, > > Did you look into these sites? > > http://directory.google.com/Top/Science/Math/Combinatorics/Software/Graph_Drawing/ > > i think this one sounds good: > > http://aragorn.ads.tuwien.ac.at/AGD/ > > please keep us informed, i'm also looking into this sort of thing. > > cya. > > Omar > > run2bmi21 schrieb: >> Thanks Alastair for the very interesting link. >> >> The kind of mapping I'd be interested in though is "static generation >> side" I think. I'm not interested in dragging objects or creating new >> vertices or edges dynamically on the SVG canvas. I'm interested in >> creating a huge static picture (a graph) of a client-server universe >> that contains about 1,500 vertices and about 5,000 edges, with none of >> the vertices hiding or overlapping each other (hidden or overlapped >> edges I don't care about), and with the vertices positioned on the 2-D >> SVG canvas roughly approximate to their position in the linear space >> of IP addresses. Linear IP address space vs. 2-D SVG canvas, minimal >> white-space, no vertex overlap: that's the problem. >> >> If I wanted to do a Google search to explore and maybe find mapping >> algorithms of this sort, what keywords would be good? Are you >> familiar with this area? >> >> Thanks again for responding, >> Doug >> >> BTW, Check out visualthesaurus.com for a really nice implementation of >> an interactive graph. The graph at >> http://jibbering.com/foaf/foafnaut.svg reminded me of it. >> >> >> --- In svg-developers@yahoogroups.com, "Alastair Fettes" >> <[EMAIL PROTECTED]> wrote: >>> Take a look at the following foaf example from Jim Ley. >>> >>> http://jibbering.com/foaf/foafnaut.svg >>> >>> Are you doing avoidance client side or static generation side? Static >>> collision avoidance is a much more complicated issue, especially if >>> you don't have the use of a nice SVG DOM. >>> >>> Alastair >>> >>> >>> --- In svg-developers@yahoogroups.com, "run2bmi21" >>> <[EMAIL PROTECTED]> wrote: One non-requirement: I don't care if the vertices partially cover the edges. The edges will be painted first, and will run from vertex to vertex as straight as the crow flys. Then the vertices will be painted on top of them. Any hints or guidance appreciated. Thanks. --- In svg-developers@yahoogroups.com, "run2bmi21" <[EMAIL PROTECTED]> wrote: > Hi, > > I was wondering whether anyone knows of a good source for mapping > algorithms. > > I'm trying to map graph vertices (which in my specific case > represent both client and server computers, each with a specific IP > address) onto a 2-D SVG canvas. The vertices are represented by > circles of varying sizes. Lines are used to depict graph edges > (which in this case represent client-server connections). > > A nice algorithm would place the vertices onto the SVG canvas > (loosely by IP address) so that: > > 1. They don't overlap > 2. They're not too crowded, i.e. they've got a bit of elbow-room > 3. They're not too widely separated (even if their IP address *are* > far apart, i.e. there's not huge amounts of white-space on the > canvas) > 4. Vertices with similar addresses are positioned close to each other > 5. Vertices with dissimilar addresses are not as close together > > I realize that these requirements can't be met 100%. I guess I'm > looking for an algorithm that might be a good compromise. > > One other feature of the graph is that it will be nearly bipartite, > a subgraph of clients with no connections between the clients, and a > subgraph of servers with a very small number of connections between > the servers, but plenty of connections between the two subgraphs. > > I'm thinking a nice arrangement might be to position "server" > vertices in the center of the SVG canvas, and "client" vertices > surrounding them. > > If you can give me any tips, thanks. > > Doug > >> >> >> >> >> - >> To unsubscribe send a message to: [EMAIL PROTECTED] >> -or- >> visit http://groups.yahoo.com/group/svg-developers and click "edit my >> membership" >> >> Yahoo! Groups Links >> >> >> >> >> >> >> > > > > - > To unsubscribe send a message to: [EMAIL PROTECTED] > -or- > visit http://groups.yahoo.com/group/svg-developers and click "edit my > membership" > > Yahoo! Groups Links > > > > > > > > - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED]
Re: [svg-developers] Re: Need algorithm ideas for mapping graph vertices onto 2-D SVG canvas
Hi, please look at this article http://www.toomuchcookies.net/archives/498/dynamic-visualization-of-network-graphs.htm and this example: http://www.toomuchcookies.net/xslt/nodes1.svg and tell me what you think. cya and good night. Omar run2bmi21 schrieb: > Thanks Alastair for the very interesting link. > > The kind of mapping I'd be interested in though is "static generation > side" I think. I'm not interested in dragging objects or creating new > vertices or edges dynamically on the SVG canvas. I'm interested in > creating a huge static picture (a graph) of a client-server universe > that contains about 1,500 vertices and about 5,000 edges, with none of > the vertices hiding or overlapping each other (hidden or overlapped > edges I don't care about), and with the vertices positioned on the 2-D > SVG canvas roughly approximate to their position in the linear space > of IP addresses. Linear IP address space vs. 2-D SVG canvas, minimal > white-space, no vertex overlap: that's the problem. > > If I wanted to do a Google search to explore and maybe find mapping > algorithms of this sort, what keywords would be good? Are you > familiar with this area? > > Thanks again for responding, > Doug > > BTW, Check out visualthesaurus.com for a really nice implementation of > an interactive graph. The graph at > http://jibbering.com/foaf/foafnaut.svg reminded me of it. > > > --- In svg-developers@yahoogroups.com, "Alastair Fettes" > <[EMAIL PROTECTED]> wrote: >> Take a look at the following foaf example from Jim Ley. >> >> http://jibbering.com/foaf/foafnaut.svg >> >> Are you doing avoidance client side or static generation side? Static >> collision avoidance is a much more complicated issue, especially if >> you don't have the use of a nice SVG DOM. >> >> Alastair >> >> >> --- In svg-developers@yahoogroups.com, "run2bmi21" >> <[EMAIL PROTECTED]> wrote: >>> One non-requirement: I don't care if the vertices partially cover >>> the edges. The edges will be painted first, and will run from >>> vertex to vertex as straight as the crow flys. Then the vertices >>> will be painted on top of them. >>> >>> Any hints or guidance appreciated. Thanks. >>> >>> --- In svg-developers@yahoogroups.com, "run2bmi21" >>> <[EMAIL PROTECTED]> wrote: Hi, I was wondering whether anyone knows of a good source for mapping algorithms. I'm trying to map graph vertices (which in my specific case represent both client and server computers, each with a specific >>> IP address) onto a 2-D SVG canvas. The vertices are represented by circles of varying sizes. Lines are used to depict graph edges (which in this case represent client-server connections). A nice algorithm would place the vertices onto the SVG canvas (loosely by IP address) so that: 1. They don't overlap 2. They're not too crowded, i.e. they've got a bit of elbow-room 3. They're not too widely separated (even if their IP address >>> *are* far apart, i.e. there's not huge amounts of white-space on the canvas) 4. Vertices with similar addresses are positioned close to each >>> other 5. Vertices with dissimilar addresses are not as close together I realize that these requirements can't be met 100%. I guess I'm looking for an algorithm that might be a good compromise. One other feature of the graph is that it will be nearly bipartite, a subgraph of clients with no connections between the clients, and >>> a subgraph of servers with a very small number of connections >>> between the servers, but plenty of connections between the two subgraphs. I'm thinking a nice arrangement might be to position "server" vertices in the center of the SVG canvas, and "client" vertices surrounding them. If you can give me any tips, thanks. Doug > > > > > > - > To unsubscribe send a message to: [EMAIL PROTECTED] > -or- > visit http://groups.yahoo.com/group/svg-developers and click "edit my > membership" > > Yahoo! Groups Links > > > > > > > - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/