Re: [brlcad-devel] Docbook synchronized with website

[Christopher S. M. <br...@ma...>]

On May 8, 2015, at 3:39 AM, Hitesh Sofat <sof...@gm...> wrote:

> I working on confluence but plugin not work. I don't have much
> knowledge about java language so its difficult to me to solve the
> error in this plugin.But i trying solve this problem and make it
> useful to me.

Did you ask anyone on their support forums for help and not get a response?

> Ok I have solution for this problem (using session technique). Don't
> worry about this  problem, it will be solved as soon as possible.

Per other message, URL simplification is a secondary issue.  Given the approach using Special completes priority #1, how would #2 and #3 be handled?

> I am working on my extension as soon as possible I am adding the
> search feature in this extension and edit also. I am trying to reduce
> the gap between them. According to me I will do this.

Hitesh, I appreciate your enthusiasm, but this is the time for exploration and discussion.  Until we’re somewhat certain that the primary considerations can be handled and we’re happy with the way they’d be handled, you shouldn’t invest a whole lot of effort and energy towards a solution we might not use.  This will be mutually frustrating, especially as you put more effort in.

There needs to be “just enough” development exploration to understand what the limitations will be.  Let’s talk through the options before you get too committed to a particular approach that might be fundamentally inadequate.  Here’s the generic process flow implied that is a cyclic loop:

A:[source files in repo compile, validate]
B:[server uploads repo content to website]
C:[website user views uploaded content]
D:[website user edits uploaded content]
E:[server enqueues website edit for review]
F:[committer reviews, commits enqueued edits]

Maybe try to characterize how a given possible solution will achieve A-F.  For example, the Confluence approach gets characterized as follows:

The Confluence Option:
A: cron job compiles checkout, stages validated docs for upload syncing
B: cron job imports validated docs via RedHat DocBook import for Confluence
C: content presented to user via Confluence wiki
D: content edited in Confluence as wiki markup
E: cron job scans log of wiki edits, exports changed content via Scroll Wiki DocBook Exporter
F: cron job stages edits into checkout for review, manually reviewed/committed by dev

Given you couldn’t get B to work, that makes this option no good.  So what other options are being considered?  If you can characterize them this succinctly, it will help decide which directions to take next.

Cheers!
Sean