INMODEL is still inefficient for large models.

The tolerances are not tight enough to accurately commpute the line length given the tension for very shallow water.

The left mouse button does not work correctly in some pick boxes (Manual Help, Contents, and Index).

The &STATUS CG_COMPARTMENT is incorrect when compartments are in moved parts.

The Strucad converter does not handle the default behavior for E and Fy correctly.

If one has panels created which are composed of triangles, each of which has an area less than 40 sq inches, then the panel will not be emitted in the &SURFACE Menu.

The concepts of -minimum and -correct are mutually exclusive. This causes problems of not being able to find equilibrium with tanks at -minimum. With this release, there was no correction of the gm, but the cg was computed correctly. This change "turns off" -correct for tanks at -minimum. It has a problem, however, in that the forces here will not be "correct" when doing a stress analysis.

No error checking was done to determine if the vertices of a panel belonged to the proper part. This adds such a check. Before if one defined a panel with vertices belonging to a different part, it created havoc with the database and any answers received were likely to be rubbish.

The way in which -SCFs are handled for beams which are not tubes is incorrect.

For large residual models (3200 degrees of freedom), MOSES will abort with a RAN OUT OF SPACE message.

If one applies large concentrated loads on both ends of a beam, there is a numerical problem in the "rigid body mechanics" part of MOSES.

In some rare circumstances, MOSES may have a segment violation. This occurs when the function heap must be resized in the middle of an append database transaction involving character data.

The MOVE command in the Connector Design Menu may not work properly if one has multiple bodies.

It may be possible to have the "Default Body" set to GROUND. This confuses MOSES.

If one defines a system, stores an event and then repeats the process several times, MOSES will loose track of the connector activity. This is a quite involved problem and this may not totally fix the problem. It is best to not do this; i.e. define all connectors before storing events.

For some situations, one can get a segment violation when doing an &instate -event. This occurs only when one of the buffers needs to be resized, but the basic data does not.

If one has a long compartment description (longer that 30 characters) then he will get a fatal error on a Windows machine then doing tank capcities.

If one computes tanks capacities in SI units with a specific gravity of 1, the weight in the tank is not the same as the volume - it is too small by 1.0002623. This is due to the conversion factor from specific gravity to density.

The use of a "power law" wind profile with a wind "table" load group produces bad results. The problem is that the height at which the wind was computed was "just above" the water surface. With this fix, the wind speed will be computed at the standard wind height (10 meters).

The command BEAM ENVELOPE -DETAIL does not work properly.

Strip Theory with small periods (1 second) will produce a numerical problem on a pc.

If one has a "large model" (24000 nodes) it is possible to get a segment violation in linux when using the command: JOINT DISPLACEMENT

Sometimes MOSES get into a recursive loop. This happens when one get an i/o error and MOSES tries to write an error message which also causes an i/o error.

A &COMPARTMENT -NO_FLOOD @ will change filling types of compartments which are not flooded.

In some situations MOSES will not open a second physical file when the base file is at the 2 GIG limit on Linux.

If one used a small number of periods, the interpolation of the wave drift force may not be correct.

On AIX 4.2 MOSES does not recognize an "end of file" condition on a sequential file.

If hydr_sum is run with a lot of periods and headings on a Windows Machine, it will abort with a floating point overflow.

Changes so that MOSES is supported on Linux on an Alpha.

Sometimes the matching for maping load does not work with selection criteria.

MOSES get confused sometimes when one only selects some parts of a problem with multiple bodies and he is computing loads. It is possible to loose all of the applies nodal loads for a part.

Several more changes to get MOSES running on Linux on Alpha. These have to do with the window interface.

This is another set of changes because of Linux on an Alpha. Basically, it does not like to do frequency domain or complex arithmetic. These changes speed things up a lot on an Alpha and a little on other machies.

When using -section on a class description to override the section properties, the specified values were included, but new section properties where not computed for use in the code check.

Sometimes, the format is not large enough for the &COMPARTMENT string function.

The option &compartment -input a b x y z does not properly override the cg of the compartment.

If someone overrides both E and G in a SACS model, it is possible to get a negative poisson's ratio.

&compartment -int_pres one xxx -open one does not work correctly.

The STRING command in the process post-processing menu is broken.

In some cases, MOSES cannot compute the correct hook location.

The weights for stiffners is, in some cases, not computed correctly.

If one tries to do g_press with 180 periods, MOSES will "run out of space while Allocating".

The previous fix "breaks" g_press for bodies which have non zero yaw when the command is issued.

The previous fix did not completely solve the problem of using a large number of frequencies and headings when computing pressures.

If one stores and event, then defines connections, and then asks for connector status, MOSES will given an error message.

If one uses more than 1 boom, then it is possible to have one of them in compression.

Overriding joint classification does not work properly.

Some "air cushion" problems may abort with a zero devide.

Using an "open valve" may produce incorrect results if the body has been "moved" with a &describe part -move command.

If one uses the "MUDMAT" command in the connector design menu for a tripod, he may get a zero divide.

The "tensioner" does not work well for an articulated stinger with 48 inch pipe in 60 meters of water.

If one changes processes and then attemps to define alias nodes, MOSES will give and error about the nodes not being defined!

In the &SURFACE menu, if one preforms certain differences and a panel has a hole in it, then the resulting block will not be closed.