This is a proof of concept for the new TIMESYS element in VOTable 1.4; the basic use case is: load some fairly randomly serialised time series, uniformise the time coordinate and concatenate them. Note: This is part of the RFC for VOTable 1.4. It is *not* part of the VOTable standard, and should be removed again some time before VOTable 1.5, or at least when sufficient support of TIMESYS is in astropy. The PoC code is in astropy_timesys.py (which ought to work with python 2 and 3 and astropy >= 1. If you want to see the input VOTables, check the foot of the file. To see the kind of code that evaluates TIMESYS metadata, see the to_jd function (though most of that is really generic conversion of various time representations to jd), and to_tcb_barycenter. WARNING: the light travel time correction from BARYCENTER to HELIOCENTER pretty certainly is wrong and needs more thought; a more general routine for light travel time correction would be highly desirable but is a bit too involved for this PoC. To see how the timeseries are merged, simply run astropy_timesys.py. This will write an essentially metadataless VOTable file joined.vot. If you scatter plot this with an aspect ratio of about 6:1, you should see a pattern... To extend the tests, look at makeseries.py, which generates the TIME_SERIES_LITERALS in astropy_timeseries (in a vi, you can place the cursor on its assignment in the source code and type ``!}python makeseries.py`` to update the time series). You'll see the time_makers, and to convince yourself that what's happening here is actually non-trivial, try changing the metadata and watch the resulting timeseries break. (Markus Demleitner ; 2019-03)