Re: IA of Max's JS files for upload?

The one thing that I think is crucial before the final import is the path
structure.
Changing paths later creates either redirects (Semantic Media Wiki has
issues with those), or links to non existent pages, which is really bad and
should be avoided.

The dom/ namespace is filled with bad paths and personally, because moving
pages creates a real mess, I do not move them and the (other) mess
continues.


☆*PhistucK*


On Sat, Jun 15, 2013 at 7:31 PM, Max Polk <maxpolk@gmail.com> wrote:

> This concerns the needed improvements to the work completed so far
> against the MSDN-JS import project, improvements like better page
> names and conversion of content to template syntax.
>
> There was a documentary about nuclear reactors that attributed the
> high cost of production in part to improvements in regulations.  By
> the time a reactor was designed and then built, such time had elapsed
> that a new set of regulations began to take effect, requiring
> considerable rework and additional costs.
>
> Such is the nature of physical construction.  Software, however, is
> very much unlike physical material.  Data input and output is more
> like soft putty to work and rework.  So rather than wait until we have
> what is perfect, my plan is to create interim versions of the work
> completed so far, uploaded to a test wiki, but also committed to
> github.
>
> The next round of tools and scripts use the interim version as input,
> and create another set of output.  Successive rounds can be created
> like this until it is complete.  The main driver is that tools and
> scripts operate on local files works much faster and more efficiently
> than against web resources (pages stored in a wiki) which have to be
> downloaded and uploaded.
>
> The spirit of a wiki is successive refinement, so having the bulk
> import effort use that pattern seems apropos.
>
> So the first round of output, all the wiki content so far plus
> upload-mapping.wiki as the mapping of where they should be uploaded,
> is now located here, which I'll use for the next round of processing:
>
>     https://github.com/maxpolk/msdn-js-conversion/tree/master/round-alice
>
> The next round will address page names.  The internal content quality
> was not reviewed yet, which would lead to a fix for empty "see also"
> sections on the bottom of some pages, removal of some Microsoft
> specific things like not using ActiveXObject that runs only in
> Internet Explorer, and so forth.
>