Started day looking for MS Excel file with statistics about employment distributed by industry: thought it might be useful for consideration of number of people needed for a small colony and the tasks each required to pursue. Eventually found the file was looking for but the statistics not quite as useful as expected.
Took another look at the conversion of my technical library to LibreOffice Basic (LoB), exported files from LibreOffice, and MS Excel and compared the basic modules with Beyond Compare, also compared with my FreeBasic version and vb.net version. The FreeBasic version seems to have been overwritten by LoB version, and needs further work to get it working again. Otherwise deleted some vba modules and renamed some Excel files, to identify the current file for further development. Used xyplorer labels to mark files as checked each one (I have coloured labels for To Do, Doing, and Done).
Rewrote VBScript which reads Basic files and extracts function headers to create FreeBasic include files (.bi). The modification was to remove the line continuation markers ( _) and write the function headers on a single line.
The other day I also rewrote VBScript I have to extract Basic modules from the LoB exported files (.xba), the original version was losing the first line of code if it was on the same line as the tag identifying start of script.
Last year I did test Git, and placed one spreadsheet into, as a test to see if this was useful for managing versions of workbooks. It probably is useful, but not all that convenient, plus I already have a ton of files for different stages of development and just straight out duplicates. I guess should do some more testing to see if it really is useful. Problem is last year I read some website to setup Git for Windows, but I cannot find similar website to see what I did: and thus what I actually have achieved by putting file in git.
Some of these things end up decreasing productivity rather than increasing. For example one place I worked had software type vault for project documents, mainly AutoCAD drawings, with drawings being checked in and checked out. All up though the system hindered people actually working on the drawings once they had been checked in: especially if someone was working on files, didn’t check them back in, and was then off sick, where upon no one could get access to the drawings to work on them. Just to name one hinderance of such systems. When the company upgraded AutoCAD they dropped the file vault system.
So I would say that version control for technical drawings is more important for technical drawings than for software development. So why is version control software such a hinderance? I’m guessing that it introduces unnatural processes, and we already have traditional document control processes, with versions (revision/issues) indicated on the documents. The problem is that the current status of a drawing may not be reflected by its current revision number, because the drawing is in process. Some offices do have in-process revision numbers for check plots: but relying on individual to revise the number and it still may not reflect the current status. If I remember one system correctly, it went something like: revision ‘A’ may be the revision number on issue, but whilst being edited it proceeds through ‘A1’ to ‘A10’ etc… Revision ‘A1’ may be floating around the office as a printed document, the CAD file on the computer may be between ‘A1’ and ‘A2’, but not yet modified and identified as ‘A2’. Whether manual or automated changing the identity code of a document every time it is changed is not very practical: and extremely wasteful of storage space.
So copying a file before modifying it is easy, and useful if need to backtrack. Using some complex system to store and retrieve the file, lacks convenience and therefore unlikely to put a file into the system, unless the changes are more significant and the needs are greater than simply backtracking.
But I will have a further look to see if git is anything more than a total git of a system.