Bugs + Fixes

Any major new software release inevitably still contains some bugs which are subsequently discovered in the field once the release has been distributed. Release 5 of the EMME/2 software is, of course, no exception to this rule. Fortunately however, the problems and bugs which were found in Release 5 since its introduction 6 months ago, are all either relatively minor or else highly specialized that only very few users are affected by them.

In November, INRO released a partial update which corrects all the bugs reported in Release 5.0, bringing up the release level to 5.2. Here is a list of some problems that were fixed:

• The network calculator, module 2.41, generated unnecessary BDISK error messages in the errors file, when used with a scenario which did not contain any transit lines.
• When printing function tabulation with module 4.14, function expressions occupying more than one line appeared garbled in the report header.
• In the auto assignment, module 5.21, the simultaneous use of link user data items ulx in volume delay functions and turn user data items upx in turn penalty functions could, under certain circumstances, cause wrong values to be used in the evaluation of the volume delay function. Since access to the upx attributes in turn penalty functions is a new feature which was introduced in Release 5, the problem is limited to new applications which were set up after the installation of the new release.
• Using a demand function index matrix in a variable demand auto assignment caused an erroneous zero demand assignment. Variable demand assignments using the same function for all O-D pairs, however, were handled correctly.
• A bug was corrected in module 6.12 which, when used with very particular link selection criteria, sometimes caused a memory access violation when plotting turning movements at intersections.
• New versions of the GPL format files for generating PostScript output, postscr.fmt and postscol.fmt, generate PostScript code which uses less stack space than before. This avoids stack overflow problems which occurred on certain printers when working with very large plot files.
• New versions of the annotation files ptfinn.ann and ptswed.ann, which contain the Finnish and Swedish versions of the customized plot titles, are provided. This corrects a problem with the special national characters that had slipped into the initial DOS versions of these files.
• Finally, the new DOS version 1.7 of EMTOOL fixes some problems that occurred with earlier versions on certain particular graphic adapters, e.g. various types of the Orchid ProDesigner board.

Release 5.2 update kits for DOS based systems of the various sizes have been sent to all EMME/2 distributors and agents. Each update kit consists of a single diskette containing all the updated modules and files and a short memo describing the update. Note that the update diskette is not signed to a specific licence, so it can be copied and installed on any EMME/2 system of the same or smaller size as indicated on the diskette. Thus, if you are using EMME/2 under DOS and are affected by any of the above mentioned problems related to Release 5.0, please contact your distributor for a copy of the Release 5.2 update kit. Updates kits for other systems are available from INRO on request.

Late breaking news: A bug in option 2 of module 2.41, "Copy attribute from other scenario", was just uncovered. It occurs when importing node, link or transit segment attributes from another scenario, if this other scenario contains more nodes, links or segments (depending on the type of attributes to be copied) than the current scenario. In this case, the attribute values of the last N elements may be copied incorrectly (where N is the difference of this type of element between the two scenarios). Until a corrected version of this module will be made available, the problem can be circumvented by using an external transfer of the corresponding attribute values (e.g. use 2.41-1 for punching first and then read the attribute values into the other scenario using 2.41-3). In some cases, e.g. for comparisons, it might also be possible to avoid the problem by performing the same operation in the "bigger" of the two scenarios.