[Pipet Devel] Re: Panic over

J.W. Bizzaro bizzaro at geoserve.net
Wed Jan 5 20:46:42 EST 2000

Brad, I'll get to all of your messages ASAP.

Brad Chapman wrote:
> Jeff, I can get two loci to break by following the following series of actions:
> 1. Add a document.
> 2. Add a converter and connect with the document. Get a good connection.
> 3. Add a processor. Connect an output (arrow) of the processor with the
> previously made connection.
> This causes the connection to break with the document and converter and
> form with the document and processor. This would be fine, except that the
> document output end still moves with the newly created link.

Ah, now I get it!  You're making a 3-way connection.  It's something I wasn't
planning on allowing in Loci, since the number of inputs and outputs permitted
for each locus must be well defined.  Plus the workflow at a junction of 3+
connectors becomes a bit confounded.  So, only 2-way connections are allowed. 
A 'splitter locus' may be developed instead to allow the workflow to be split
or duplicated to run along more than one path.

I'll fix it.  Thanks for the bug report.

>         I also noticed another thing when I was messing around. If you move
> a dot directly onto a locus (ie. on top of the two people) you lose this
> input or output. Just something that will have to be worked around
> eventually so dumb people like me don't keep accidentally losing their
> connecting dots.

Oh yeah, I noticed that a while back.  Basically, the connector can be dragged
under the symbol of its own icon, and when it's dropped, there's no way to get
it out.  The solution I was planning on was to define a radius that clears all
locus ornaments.  And when an unconnected connector is dropped within that
radius, the connector is automatically extended to fall beyond it.

>         BTW, I really like the "windowlet" idea for the dots. This way you
> wouldn't have to worry about trying to write a whole bunch of info inside a
> dot and ending up with dots that are as big as the loci! I guess this will
> have to work by getting the info from an XML document. Hmmmm...

Yep.  Connector information is kept as XML along with some other stuff about
the locus.

                      |           J.W. Bizzaro           |
                      |                                  |
                      | http://bioinformatics.org/~jeff/ |
                      |                                  |
                      |           THE OPEN LAB           |
                      |    Open Source Bioinformatics    |
                      |                                  |
                      |    http://bioinformatics.org/    |

More information about the Pipet-Devel mailing list