This is working copy of the famous function isfloat. This one was building entirely in X86 with Intel syntax. The source file like most assembly source files appears highly confusing. This isfloat source file is accompanied by an execution flow chart that was created before the isfloat function was programmed. The algorithm was first debugged in the flow chart and then translated to X86 Intel source code. No further debugging of the algorithm was required. The only fixes required by the source code was to carefully selection which size of registers to use: one-byte registers, 8-byte registers, or some other size of register.
The following are available for download. Know that isfloat is a library function, as such, it is not part of any one application. Consequently, there are no accompanying software files with isfloat.
Isfloat-execution-flowchart.odt <== Best viewed with LibreOffice after downloading rather than view in Google docs.