On Mon, Jan 19, 2015 at 03:03:21AM -0700, NopMap wrote:
> On the other hand, just reverting them does not feel right to me either.
> Some of the examples have their merit.
>
> What do you think?
so lets look at the points that were reverted:
##== Disadvantages of semicolon separated lists ==
##* parsing of values is required
sure parsing is required. How terribly difficult is it to split
a string by ";"?
##* uncertainty if order is important: {{tag|example|value1;value2}} vs
{{tag|example|value2;value1}}. Should data consumers threat them differently?
Very few ''mappers who use semicolon'' care to define this.
one of the reasons of the existence of multivalued entries is that the
mapper could not decide what is more important.
Data consumers are not required to be smarter than the mapper but specialised
consumers could pick the value they care about.
##* insane complexity when using simple querying tools ([[JOSM/Search
function|JOSM search]], [[OverpassQL]]) <!-- can't think about good example ATM
-->
is it really so hard? In most cases not even regexps are required.
##* prone to errors and they are hard to detect without parsing of values. In
contrast to this, if you make mistake in [[key]] it will be easily detected by
users of [[Taginfo]] and we may correct tags planetwide.
I would be curious to know if it would be hard to change taginfo to offer
alternative views to multivalued entries.
Richard
_______________________________________________
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging
