Page 1 of 1

Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 9:23 am
by hszhao.cn@gmail.com
I noticed the following example here:
AFM
2.80000
1.00000 .00000 .00000
.00000 1.00000 .00000
.00000 .00000 1.00000
1 1
Carthesian
.00000 .00000 .00000
.50000 .50000 .50000
But I'm puzzled by the variable name Carthesian used here, more specifically, I think it should be written as Cartesian. For validation, I also checked the vasp source code as follows:

Code: Select all

werner@X10DAi:~/Public/hpc/vasp/release/vasp.6.3.0$ ug Cartesian |wc
    228    1292   23085
werner@X10DAi:~/Public/hpc/vasp/release/vasp.6.3.0$ ug Carthesian |wc
      9      58     568
As you can see, there are a total of 228 instances of Cartesian and 9 instances of Carthesian. Based on this result, it seems that both versions are reasonable, but Cartesian is preferable. Am I right?

Regards,
Zhao

Re: Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 9:57 am
by alexey.tal
Dear Zhao,

Thank you for noticing this typo. However, it should not affect the calculation as the poscar reader uses only the first character, so both spelling variants yield identical results.
I have corrected the wiki page.

Re: Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 1:08 pm
by hszhao.cn@gmail.com
However, it should not affect the calculation as the poscar reader uses only the first character, so both spelling variants yield identical results.
If so, how does the reader/parser discern two tags with the same first character?

Re: Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 1:12 pm
by alexey.tal
This doesn't concern INCAR parser. The tags in the INCAR files have to be spelled correctly to be "understood" by the code.

Re: Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 1:20 pm
by hszhao.cn@gmail.com
According to your description above and our discussion, it seems that these tags in the POSCAR file only serve as a data block separator. Is that right?

Re: Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 1:25 pm
by alexey.tal
Since there are essentially two kinds of coordinates "Cartesian" and "Direct", only the first character on the line is significant and the only key characters recognized are C, c, K or k for switching to the "Cartesian" mode. Everything else will be interpreted as "Direct" mode.

Re: Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 1:37 pm
by hszhao.cn@gmail.com
However, the prerequisite for doing so is that the corresponding data block must be correctly identified as coordinates in this case. Because there are many types of data blocks in VASP, I am not sure whether relying solely on the first character of the tag to distinguish them is sufficient for all cases.

Re: Cartesian vs Carthesian

Posted: Thu Aug 10, 2023 1:44 pm
by alexey.tal
Correct. This data block must be identified as coordinates. We have a detailed description of the POSCAR file format on the wiki. Please let me know if you think that some relevant information is missing there.