# [rfc-i] Proposed way forwards on backward compatibility with v2

Riccardo Bernardini framefritti at gmail.com
Tue Feb 18 03:53:37 PST 2014

On Mon, Feb 17, 2014 at 10:26 PM, Dave Crocker <dhc at dcrocker.net> wrote:
> On 2/17/2014 1:11 PM, Julian Reschke wrote:
>>
>> On 2014-02-17 21:50, Dave Crocker wrote:
>>>
>>> And in this case it isn't appropriate to try.  There is nothing 'broken'
>>> in the current version.  There is merely the desire for some additional
>>> features.
>>
>>
>> There are certain things we'll want to discourage, because they are
>> workarounds for missing stuff. Whether we'll be able to remove those is
>> a separate question (I hope we can if there's sufficient time between
>> deprecation and removal).
>
>
>
> Yes, but...
>
> This entire discussions is an instance of a classic debate between theory
> and practice that has a particularly long history for document
> representation/presentation.  The debate is between being purely
> 'structural' in constructs -- with absolutely all presentation details being
> left to the printing/formatting engine later -- versus permitting explicit
> presentation details in line.  <p> is an example of the former, and <vspace>
> is the latter.
>
> A purist structural model is extremely appealing.  The problem is that it is
> also Procrustean.  Some of you might like having your legs get chopped off
> but some of you might not...
>
> So yes, it's reasonable to press people to use the cleaner, more abstract
> construct.
>
> But no, it is not reasonable to remove all ability to overcome perceived
> limitations with the abstractions.
>
> Rather, we need to understand that the abstraction approach is unlikely ever
> to be perfect, and that the indelicate pragmatics of direct formatting
> directives is likely to continue to be (sometimes) useful.

Let me strengthen this with my experience.  I actually love the
"abstract" approach an 99.9% of the times LaTeX does a wonderful work
about how the final document looks (if you use the right tools for the
right work, otherwise you can get mad).  LaTeX is not perfect,
however, so there is a 0.1% of the cases where you need to do some
adjustment with low-level commands like \hspace, \vspace and so on.

So, +1 to promoting the use the abstract structure, but keeping
low-level commands for the occasional "hand-hacking".