Fortran use in Industry

[JohnNichols]

As was pointed out when I started playing with the prestress program from TxDOT, there is a current version of the program on their website. 

After I had the program running and it appears to give reasonable engineering answers, I downloaded the txdot program and checked the answers for some of the sample problems.   

The EXE file shows a compile date of 2016.  I cannot insert the picture into the post.  Even if I borrow a coffee mug.  The file shows it was compiled from Intel Visual Fortran, clearly from a version dated 2016 or earlier.  

$        11111111112222222222333333333344444444445555555555666666666677777777778
$2345678901234567890123456789012345678901234567890123456789012345678901234567890
PSF No    County---------     Highway---     XXXX-XX-XXX  CodedBy
Example 1: LRFD Spec, 110 ft Span, Type IV, 8.5' Bm Spacing, 8.0" Slab
PROB Standard I-Beam using 2004 LRFD Specs
$MAT1        MoE                           MoE
$MAT1        of                            of
$MAT1        Beam                          Slab
MAT1         5000.0                        5000.0
$OUTP   LF - LF outputs tables of moments, shears, and stresses at tenth points
$OUTP   SF - SF is default and it omits the listed LF tables
OUTP    LF 
$SPEC    X,  where: 0=95/02 Stnd Spec(default), 1=94 Stnd Spec,
$SPEC    |          5=2004-2007 LRFD, 9=2009-2010 LRFD, A=2012 LRFD   
$SPEC    |Y, where: 0=1994 losses, 4=2004 losses, 7=2007 losses
$SPEC    ||         A=TxDOT 2012 losses 
$SPEC    |(Design Spec default= 0 for SS & AREMA)
$SPEC    ||(Losses Spec default= 4 for LRFD)
SPEC     54
$LLDF     Shear
$LLDF     LLDF
LLDF      0.849
$BEAMSpan Beam Beam           Beam Slab Comp Mom                      UDL on 
$    LabelLabelType SpanLengthSpac.ThickSlab LLDF    RH               Comp Sec
BEAM ALL  1    IV   108.58    8.500 8.0      0.696   60               0.132

The input file is like an old card deck. The last line is almost identical to the 1970's input line for the beams and it uses the same coding IV, for the beam type.  

The notes say it was based on a 1990 version of the prestress program,  but points to the 1970 Fortran version.  

The manual talks of each line of the input file as a card.   LF and SF reflect the 1970 version that had long form and short form for the output.  Some of the output titles are identical to the 1970 version. 

The results would indicate that they are still using a lot of the 1970s code for the math.  

Interesting.  The manual shows similar pictures to the 1970s manual for the cards.  

[cryptogram]

I started my career working on a staff that created a simulation model of a reservoir system,  and water temperature and water quality models of rivers and reservoir.   This all started out on cards using mainframe computers, and everything was in Fortran.

These models are still in use today.  One application involves using the simulation model to create a long term forecast of boundary condition flows, and then running various 1 and 2d river and reservoir models to forecast temperatures throughout the reservoir system.

The Fortran code is mostly unchanged, just switched to using files for inputs and output.   Everything runs from a VB.net controller that

talks to database and web services, builds input files, runs the Fortran models, reads model results and stores the results in databases.

I actually retired from full-time work 6 years ago, but still work as a part-time contractor supporting various legacy systems.  I'm the last remaining member of the original model development staff, trying to train up someone to support these models.   Looks like I have a meeting scheduled next week, where he will find out that he is supporting Fortran code.