Inspired by discussion in #619 I came to realization that there is no consistency ATM: .tsv "Text" examples come both in
proper TSV form
$> git grep -e ' ' | grep '\.md'
src/02-common-principles.md:onset duration response_time correct stop_trial go_trial
src/02-common-principles.md:200 200 0 n/a n/a n/a
src/03-modality-agnostic-files.md:filename acq_time
src/03-modality-agnostic-files.md:func/sub-control01_task-nback_bold.nii.gz 1877-06-15T13:45:30
src/03-modality-agnostic-files.md:func/sub-control01_task-motor_bold.nii.gz 1877-06-15T13:55:33
src/04-modality-specific-files/08-genetic-descriptor.md:participant_id age sex group genetic_id idh_mutation
src/04-modality-specific-files/08-genetic-descriptor.md:sub-control01 34 M control 124587 yes
src/04-modality-specific-files/08-genetic-descriptor.md:sub-control02 12 F control 548936 yes
src/04-modality-specific-files/08-genetic-descriptor.md:sub-patient01 33 F patient 489634 no
src/05-derivatives/03-imaging.md: anat/
src/05-derivatives/03-imaging.md: sub-001_space-orig_probseg.nii.gz
src/05-derivatives/03-imaging.md: sub-001_space-orig_probseg.json
and more often in some shape of "replace tab with some spaces" - some times with nice layout into a tabular form, some times not really (even canonical example had no tabs but varying number of spaces used for delimiting,
#643 ). IMHO ideally
- markdown sources should use TABs where TABs are intended to be used
- rendering into HTML or PDF should do its best to signal that they are tabs while providing visually coherent display of the table.
I.e. ideally it should be a job of remake (some plugin) to do use TABS (ATM they are replaced by spaces by remake I guess), and just use HTML/CSS to provide nice visual alignment into the table (should be possible) while having TAB as the actual character so content could be copy pasted. Even though it is unlikely the actual entries would be copy pasted, I would expect some readers copy-paste headers (which also should be tab-delimited). Could even use (although not really needed and overkill) smth like https://www.datatables.net/, where copy pasting a row from that fancy table does produce tab delimited entries:
$> xsel -o | sed -e 's, ,·,g' -e 's,\t,⇰,g'
Name⇰Position⇰Office⇰Age⇰%
Inspired by discussion in #619 I came to realization that there is no consistency ATM: .tsv "Text" examples come both in
proper TSV form
I.e. ideally it should be a job of remake (some plugin) to do use TABS (ATM they are replaced by spaces by remake I guess), and just use HTML/CSS to provide nice visual alignment into the table (should be possible) while having TAB as the actual character so content could be copy pasted. Even though it is unlikely the actual entries would be copy pasted, I would expect some readers copy-paste headers (which also should be tab-delimited). Could even use (although not really needed and overkill) smth like https://www.datatables.net/, where copy pasting a row from that fancy table does produce tab delimited entries: