Jupp53 wrote on 04/24/20 at 18:03:15:

The first part I competely agree with. And ChessBase hasn't fixed the known issues in forever. I mostly use CB for opening CBV files, and for converting CBH files to PGN.

The second part I partially agree with. CA, unlike CB, handles the headers well. But move-text not so well. Exporting a Chess Assistant format database to PGN results in a lot of CA-specific "%" comments that other programs don't know what to do with, as well as some other non-standard things that are problematic, e.g. multiple comments like {1}{2} attached to a single move.

I regularly use three different chess database GUIs (Shredder, Chess Assistant, Tarrasch), as well as many other chess programs (pgn-extract, EPD2diag, Arena, XBoard and Winboard, Bookup, ChessBase, Fritz, etc.). Each program is stronger at some feature, and I'm frequently trying out new ones to see if there is something even better for some purposes. So PGN and EPD are critical for me, as they are the only formats all programs understand.

Despite there being published standards for both, there are big problems with interoperability.

Potential issues with PGN are many. Off the top of my head:

• 32-bit programs can't read a file bigger than 2 GB.
• The pgn standard says any line ending is acceptable, but some programs that expect Windows CR-LF endings can't read files with Unix LF or Mac CR endings.
• The pgn-standard says only plain ASCII encoding, but these days most programs import and export UTF-8. So maybe ASCII is needed, or maybe Windows-1252, or maybe a UTF-8 byte-order mark is needed, or maybe some higher characters are fine but others are problematic.
• Chess-960 games are finding their way into databases, but not all programs handle those.

First thing I do with a new PGN file is open it in a decent text editor, to see if it looks okay.

Second thing I do is run it through pgn-extract. It's a life-saver!
https://www.cs.kent.ac.uk/people/staff/djb/pgn-extract/
$ pgn-extract -r -s myfile.pgn

A good trouble-shooting step is exporting a fairly small database from SCID, and seeing if CB Reader 2017 can import that successfully.

Edited: