Due to a variable not initialized as it should be, the Variance computed by the clustering coefficient algorithm is correct only on the first computation.
When the user executes the algorithm repeatedly, the variance is not initialized again thus the algorithm reports increasing and wrong variance values.

When the user selects/embeds a layout model in the network, SocNetV does the work but it does not display anywhere the layout name somewhere. Thus there is no way for the user to remember later what layout she has applied.
SocNetV should display current layout somewhere, either in a new text box or on the canvas.

When isolated actors exist in a network (and are not disabled) the Eccentricity (max geodesic distance to all other actors) of each actor is by default infinite.
The Eccentricity Report states that fact although it reports non-inf scores for all other actors (except the isolated) as if the isolated one was disabled by the user. This is misleading.

SocNetV needs to support the Graph Reciprocity metric.
In plain words, the reciprocity of a graph is the fraction of reciprocal edges over all edges of it.

To be more exact, SocNetV should support both Arc and Dyad Reciprocity measures:

• The arc reciprocity is the fraction of reciprocated ties over all present ties of the graph.
• The dyad reciprocity is the fraction of actor pairs that have reciprocated ties over all connected.

References:
https://en.wikipedia.org/wiki/Reciprocity_(network_science)
https://reference.wolfram.com/language/ref/GraphReciprocity.html
http://www.analytictech.com/ucinet/help/hs4345.htm

SocNetV should implement Bonacich Power Centrality BPC (alpha,beta) (not to be confused with the already implemented Power Centrality (PC), a generalized degree centrality measure suggested by Gil and Schmidt).

See references
Bonacich, P. (1972). “Factoring and Weighting Approaches to Status Scores and Clique Identification.” Journal of Mathematical Sociology, 2, 113-120.

Bonacich, P. (1987). “Power and Centrality: A Family of Measures.” American Journal of Sociology, 92, 1170-1182.

At the moment SocNetv supports this set of commands:

graph
comment
directed
node
edge
id
label (inside "node")
label (inside "edge")
source
target

See Technical Report here:
http://www.fim.uni-passau.de/fileadmin/files/lehrstuhl/brandenburg/projekte/gml/gml-technical-report.pdf

A quite large set of test files can be reached here:
http://www-personal.umich.edu/~mejn/netdata/

At the moment, progress dialogs are not intuitive and do not always show the exact progress the app has made in the algorithm it executes. SocNetV should have proper progress dialogs, each one of them depicting the process of each one algorithm.

It would be nice if the user could add arbitrary text anywhere on the canvas.
The text (multi-line or not) could act as remarks/notes of the user on various parts of the network.

Due to a typo in computing Group Degree Centralisation (versions prior to 2.2), the GDC score of a network was always reported to be divided by N-1.

SocNetV uses qprogressdialogs to inform the user about the progress of a running algorithm/job.
These dialogs have a Cancel button to cancel the job, but pressing it does really nothing leaving the frustrated user waiting for the job to finish.

SocNetV should compute Network Connectivity.
This is the number of nodes that would have to be removed in order for one actor to no longer be able to reach another. If there are many different pathways that connect two actors, they have high "connectivity" in the sense that there are multiple ways for a signal to reach from one to the other.

SocNetV 2.2 for Linux writes a lot of characters (megabytes) into /var/log/syslog when doing an analysis.
But I'm not sure how to reproduce it.

I find this in syslog:
Mar 8 21:43:15 mymachine compiz[2985]: x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x

Press F7, recreat Krackhardt's high tech managers. The loaded network is directed, not symmetric.
Analyze Cohesion > Connectedness.
The result popup says:
"This undirected graph is not connected."

SocNetV has a built-in web crawler, which atm is very simplistic and produces something like ego-social networks:
It scans a 'source' node/webpage for "a href" links to other webpages (which become nodes in the social network) and then visit and scan every one of these other webpages for href links to other pages and so on.

ATM, the only option provided to user is whether the crawler should scan only internal urls (pages in the same domain), external urls (pages in other domains) or both.

The crawler should be enhanced with useful options such as:

• domains whitelist: allow the user to indicate which domains should be scanned.
• domains blacklist: allow the user to exclude some domains in bulk.

1. Create a .2sm file with the Davis, Gardner and Gardner dataset

1 1 1 1 1 1 0 1 1 0 0 0 0 0
1 1 1 0 1 1 1 1 0 0 0 0 0 0
0 1 1 1 1 1 1 1 1 0 0 0 0 0
1 0 1 1 1 1 1 1 0 0 0 0 0 0
0 0 1 1 1 0 1 0 0 0 0 0 0 0
0 0 1 0 1 1 0 1 0 0 0 0 0 0
0 0 0 0 1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 1 0 1 1 0 0 0 0 0
0 0 0 0 1 0 1 1 1 0 0 0 0 0
0 0 0 0 0 0 1 1 1 0 0 1 0 0
0 0 0 0 0 0 0 1 1 1 0 1 0 0
0 0 0 0 0 0 0 1 1 1 0 1 1 1
0 0 0 0 0 0 1 1 1 1 0 1 1 1
0 0 0 0 0 1 1 0 1 1 1 1 1 1
0 0 0 0 0 0 1 1 0 1 1 1 1 1
0 0 0 0 0 0 0 1 1 1 0 1 0 0
0 0 0 0 0 0 0 0 1 0 1 0 0 0
0 0 0 0 0 0 0 0 1 0 1 0 0 0

2. Import the file in SocNetV. The network created does not make any sense from the user perspective. In fact, the network which is drawn upon the canvas is the acquaintance network by constructing a 1-mode matrix A such that aij = 1 if person i is in at least one group with person j.

Furthermore the two_sm_mode variable is not used by Parser.

SocNetV should be able to import two-mode data by:
a) creating bipartite graph with both modes (groups of actors/organizations)
b) creating the acquaintance network (as it does now).

The app crashes if a search (Ctrl+F) is done after removing a node.
To reproduce:

1. Open a net (i.e. "Countries_Trade_Basic_Manufactured_Goods" by Wasserman-Faust)
2. Ctrl+F search for a node.
3. remove that node (Ctrl+Alt + . )
4. Press Ctrl+F again. SocNetV crashes.

Everytime I try to open a Pajek formatted file, including my own or the one provided in SocNetv manual, I'm getting an error message "XML tokenstring at line 1 is invalid".

SocNetV craches when doing Hierarchical Cluster Analysis in networks with isolated nodes. This happens only when the input matrix is the geodesic distances between actors.
To reproduce:

1. Recreate Padgett's Florentine Families network (press F7 then select it)
2. Perform HCA on that network from the menu Analyze -> Structural Equivalence -> Hierarchical Clustering. Leave default options (input matrix: distances)

Instead of strange combinations Ctrl+. and Ctrl+/ for new node and edge creation, we should use INSERT (or ENTER) key for node creation and / for edge creation.
Moreover we can automate the creation of a new node connected to an existing node N, by the combination SHIFT+INSERT iff a node N is already clicked/selected.

To delete a node or and edge, it should be suffice to click on it and press key: DELETE - For that, we need to have a delete action which will check if a node or an edge is clicked and then trigger the proper node or edge removal from canvas.

As the tool is properly using static networks, but it will be better if the tool can incorporate dynamic networks into it.

SocNetV should implement Katz Centrality to enable the user to measure the degree of influence of an actor in a social network by taking into account the total number of walks between a pair of actors.

Katz centrality computes the relative influence of a node within a network by measuring the number of the immediate neighbors (first degree nodes) and also all other nodes in the network that connect to the node under consideration through these immediate neighbors. Connections made with distant neighbors are, however, penalized by an attenuation factor \alpha. Each path or connection between a pair of nodes is assigned a weight determined by \alpha and the distance between nodes as \alpha^d.

For example, in the figure on the right, assuming that John’s centrality is being measured and that \alpha = 0.5. The weight assigned to each link that connects John with his immediate neighbors Jane and Bob will be (0.5)^1 = 0.5. Jose connects to John indirectly through Bob, the weight assigned to this connection (composed of two links) will be (0.5)^2 = 0.25. Similarly, weight assigned to the connection between Agneta and John will be (0.5)^3 = 0.125.

Mathematical formulation

Let A be the adjacency matrix of a network under consideration. Elements (a_{ij}) of A are variables that take a value 1 if a node i is connected to node j and 0 otherwise. The powers of A indicate the presence (or absence) of links between two nodes through intermediaries. For instance, in matrix A^3, if element (a_{2,12}) = 1, it indicates that node 2 and node 12 are connected through some first and second degree neighbors of node 2. If C_{\mathrm{Katz}}(i) denotes Katz centrality of a node i, then mathematically:

C_{\mathrm{Katz}}(i) = \sum_{k=1}^{\infin}\sum_{j=1}^n \alpha^k (A^k)_{ji}

Note that the above definition uses the fact that the element at location (i,j) of the adjacency matrix A raised to the power k (i.e. A^k) reflects the total number of k degree connections between nodes i and j. The value of the attenuation factor α has to be chosen such that it is smaller than the reciprocal of the absolute value of the largest eigenvalue of the adjacency matrix A.[4] In this case the following expression can be used to calculate Katz centrality:

\overrightarrow{C}_{\mathrm{Katz}} = ((I - \alpha A^T)^{-1}-I)\overrightarrow{I}

Here I is the identity matrix, \overrightarrow{I} is an identity vector of size n (n is the number of nodes) consisting of ones. A^T denotes the transposed matrix of A and (I - \alpha A^T)^{-1} denotes matrix inversion of the term (I - \alpha A^T).

Urls:
http://en.wikipedia.org/wiki/Katz_centrality

SocNetV crashes when the user performs Information Centrality after he has disabled isolated nodes.
Reproduce:

1. Open Padgett Florentine families.
2. Disable isolated node (Ctrl+X, Ctrl+F)
3. Run Information Centrality (Ctrl+8)

For increased performance, SocNetV should use as many cpu cores as possible as well as use minimum memory.

ATM, SocNetV runs pretty much in the same cpu core. Only file parser and crawler functions are using differrent cpu threads. Also, when the network has hundreds of nodes and the user moves some of them, there is a constant increase in memory usage.

After almost 10 years in sourceforge vcs (svn, then git) and a brief experiment with bzr, it's time to move the project code development in a reliable and trusted git hosting service for everybody to see and contribute. I am happy to report that this service will be Github! The codebase (branches, etc) has been imported painlessly and we have been developing version 2.2 using github git for the past month without any problem. So, from now SocNetV will use GitHub.

Save this network to some file directed.dot:

graph graphname {
// This attribute applies to the graph itself
size="1,1";
// The label attribute can be used to change the label of a node
a [label="Foo"];
// Here, the node shape is changed.
b [shape=box];
// These edges both have different line properties
a -- b -- c [color=blue];
b -- d [style=dotted];
}

2. Load directed.dot in SocNetV. All network edges appear blue. The default edgecolor should be changed to blue.

3. Create a new edge by mouse (middle/double clicking). The new edge takes the previously set initEdgeColor (default: black) rather than the color set by the network file.

A user wrote:
"My main frustration, as a beginner, is that I can’t find an undo button; when I make large-scale mistakes, I find myself having to start all over, and that’s quite frustrating. "

SocNetV needs to have undo functionality to assist users drawing networks on the canvas.

The matrix output shows extra columns (one for each disabled node).

SocNetV needs to provide more methods for the identification of cohesive subgroups in one-mode networks.
-n-cliques
-n-clans and n-clubs
-k-plexes

• matrix permutation approaches

Due to a variable not initialized properly, in some cases (i.e. plotting adjacency matrix) the informational message displayed in the progress dialog is not updated but instead it shows the same text from the last user action involving progress dialogs.

SocNetV supports Eigenvector Centrality for analysis and layout, but there is not EVC options in Layout menu (either in radial or level type).

When the user changes the negative edge colour in Settings, a bug in the code changes the positive edge colours as well. Also, if the user changes the negative edge colours, the settings.conf is saved with wrong colour values for future sessions.

GraphViz (dot) files can't be parsed if the data is in a single line.

For instance this graph is read but nothing is displayed:
digraph mydot { node [color=red, shape=box]; a -> b -> c ->d node [color=pink, shape=circle]; d->e->a->f->j->k->l->o [weight=1, color=black]; }

If we use line breaks, then everything is OK. Test:

digraph mydot {
node [color=red, shape=box];
a -> b -> c ->d
node [color=pink, shape=circle];
d->e->a->f->j->k->l->o
[weight=1, color=black];
}

Except HCA, SocNetV should implement the following clustering algorithms to automate finding community structure:

• the Girvan-Newman algorithm, older one for small graphs mainly, see paper " Community structure in social and biological networks"
• the Clauset-Newman-Moore algorithm, see paper "Finding community structure in very large networks"
• the Wakita and Tsurumi algorithm, see paper “Finding Community Structure in Mega-scale Social Networks”

Nodes "marked" by user (ie. via search) should be maintained in Graph, not in MW ui.

A user suggests:

It would be nice if SocNetV could apply a layout where the color of each node indicates its clustering coefficient.

ATM, SocNetV supports two layout methods:
(a) force-directed placement (FDP)
(b) radial / level by prominence index score

The FDP layout algorithms in SocNetV have been somewhat improved, but we need to add more FDP models, i.e. Kamada-Kawai

SocNetV needs to provide methods for the identification of cohesive subgroups in one-mode networks.
-n-cliques
-cores
-n-clans and n-clubs
-k-plexes

• matrix permutation approaches

Due to a variable not initialized properly, the connectedness method reports wrong results when executed after an algorithm related to distances. For instance open a connected network, run distances matrix, then connectedness. SocNetV will report it as unconnected.

Create a test net of 3 actors.
0 1 0
1 0 0
0 0 0

Select Edit > Filter > Disable Isolate Nodes. The actor 3 is disabled and hidden.
But when you do Analyze > Cohesion > Distances matrix the reported matrix is still a 3x3 matrix not a 2x2 one as it should be.

By default the edge arrow size follow the width of the edge line.
SocNetV needs to allow the user to control arrow size.

Reports can be more helpful if they offer scatter plots of the results. For instance, a centrality report might include a scatter plot of actor centralities. This would be a handy tool for the researcher to overview the results, as well as another visualization per index layout -- only this time inside the report.

SocNetV supports prominence-based layouts where the size of all nodes follows the prominence score of each node. Nevertheless, this functionality is only available through the toolbox. There are no relevant options in the Layout menu, except Node Size by inDegree or outDegree.

"I want to know if alternative icons can be substituted for nodes. I have both *.ico and *.png icons that I have made specific to my work. I would like to use those. Is this possible with socnetv?"

A regression in v2.2 left the Control Panel layout by closeness centrality index option (radial, level, nodal) not working. The user selects Closeness Centr index, clicks Apply but nothing happens. Nevertheless, the same option in the main menu works fine.

The reciprocity of a graph g is the fraction of reciprocal edges over all edges of g.

see https://reference.wolfram.com/language/ref/GraphReciprocity.html

Valued edges or arcs with negative weights in [-1,0) are not displayed at all.
Different negative weights are displayed properly (with dashed lines as expected).

SocNetV omits zero-valued edges during network analysis computations. Nevertheless, when opening weighted edge lists, SocNetV accepts edges with zero values/weights and draws them (although completely by mistake).

A SocNetV user informed me that they need this functionality (zero valued edges to be drawn) and on top of that to be able to select some color for this kind of edges.

We will implement this functionality in Settings but it will be explicitly declared that the functionality serves only for those that have data (i.e. weighted edge lists) with zero weights on some edges and they still need to draw those edges, despite the fact that they will not be considered in any SNA computation by SocNetV.

Note: This functionality needs investigation as to what formats (except edge lists) support zero-valued edges.

SocNetV up to 2.3 computes wrong Betweeness Centrality scores when the loaded network is weighted.

For example, take the 4 actors social network:

0 | 2 | 0 | 2
2 | 0 | 1 | 0
0 | 1 | 0 | 1
2 | 0 | 1 | 0

Apparently, each of actors 2 and 4 lie in half the shortest paths (geodesics) from 1 to 3 . Thus their BC should be 0.5
And actor 4 lies on the sole geodesic between 2 and 4, this its BC should be 1.
But SocNetV reports zero BC for all actors except 4.

SocNetV should be able to visualize nodes'prominence by using hue (from red to blue)

See
http://en.wikipedia.org/wiki/Centrality#mediaviewer/File:Graph_betweenness.svg

Testing really large files (see example net in #23) showed that performance options should be added in Settings. For instance, when the loaded network has thousands edges as in your data, an option to "Disable Edge Highlighting when hovering" would be handy, since that would speed up user interactions with crowded nets.