[Owasp-leaders] Proposal for changing OWASP Documentation

Gary Robinson gary.robinson at owasp.org
Sun Feb 14 20:30:44 UTC 2016


Hi Folks,



I want to reach out to the leaders and bring up the subject of OWASP docs
projects, in fact all OWASP documentation including wiki and text available
to our code projects as well.  I’ll try to keep this e-mail brief so won’t
get bogged in details, but generally I want to know if the community
experiences the same pains, see the same opportunities, or has other
suggestions for improvement.



The issues that can arise from our current method of developing our various
docs include:



   1. Draft content that exists in the wiki.  This may be in varying states
   (correct, incorrect, lousy, confused, etc.) and is visible to the internet
   and typically not clearly labelled as draft.  Google ‘owasp purple monkey
   dishwasher’ for an example of a draft wiki page visible to the internet.
   This content also needs to get cleaned up after a project release.
   2. Substandard descriptions/content can get into our docs.  Getting
   people to review every line/example/diagram/appendix is difficult with a
   volunteer organization (as other threads have discussed)
   3. Duplications happen, as 10 different projects create/copy/paste their
   definitions of topics such as XSS, SQLi, CSRF, etc.  This wastes effort in
   an organization already constrained of active volunteers.
   4. Content gets out-of-date.  The work to create a new version of a doc
   project takes years.

I’m sure some readers will be mentally adding their own issues to this list.



Proposal:


Bringing together discussions with a few people over the last few years
(you know who you are), I’m proposing the following: we write our docs with
reusable resources.  What would this mean?  Something similar to the
following:



   1. We dump all of the content from our wiki, current docs, descriptions
   in code tools, etc.  We put it into markup (as some projects are already
   doing) and add it to source code repositories.
   2. We share doc markup files *across ALL docs and code projects*.  For
   example, imagine we have a folder for SQLi.  This directory contains the
   OWASP ‘golden source’ for SQLi definition, examples, code, tests, etc.
   Repeat for all other AppSec issues (CSRF, cert pinning, etc.).  We use a
   mechanism to ‘compile’ these markdown files into PDFs and integrate into
   code project HTML pages.
   3. Similar to good coding projects, we control who can edit the files
   under certain directories – people we know have expertise in an area.
   Edits get peer reviewed before submission.  Other people can suggest edits
   and prove their experience to the existing team to join it.
   4. We allow anyone to ‘include’ this markup file into their project.  So
   if the Code Review Guide wants to add a section on SQLi, and needs a
   definition, I don’t write it (or copy from wiki), I simply include the
   relevant markup file.  Same for testing guide, dev guide, ZAP hints page,
   security shepherd info page, cheetsheet, and on and on.
   5. We allow all of our docs, plus the wiki, plus all code projects, to
   dynamically use an markup file update.  We make this ‘real time’.  This
   needs an example.  Say in March a massive change occurs in the world of
   SQLi.  Right now any project that talks about SQLi would need to manually
   go in and update, and those updates will be of varying quality and
   content.  If, instead, one (true) source file was update, all those other
   projects could spot the change and automatically rebuild themselves,
   meaning the next person to download a development guide PDF, or view the
   wiki, would get the updated SQLi information.
      1. This is a big change.  This may be a controversial change.  *However
      it would greatly reduce our workload* (only one markup document needs
      to get updated).  It will also *greatly reduce review tasks*, as
      everyone is sharing core content which is reviewed once.  It
also improves
      our image to the world, as all projects have the same great descriptions
      and content.
      2.   This change also improves our responsiveness.  Imagine a
      heartbleed type issue being reflected in all OWASP code and documentation
      projects, as well as the wiki/cheetsheets, within a few days?
(simply the
      time for the team to agree updates to the text/examples/descriptions,
      review, and submit)
   6.  We should also make these markup files available to anyone on the
   internet (read only).  This way the source descriptions become an OWASP
   resource it itself, and anyone out there needing to spread the word on
   AppSec has easy access to rock solid, up-to-date definitions.

This changes the model, from people like myself who run ‘projects’, to
smaller expert teams who know ‘technologies’ (such as SQLi or IIS secure
configuration).  It focuses people where they want to be on docs projects,
but easily shares that knowledge across all OWASP (and more) projects.  It
also means there’d never be another need to clean-up the wiki – it would
always be based off the markup content.


*Big downside*: there’s a large piece of work to start it off.  All content
would need to get organized, put into sensible structure, converted to
markdown, argued over, ‘experts’ defined and assigned, etc.  I doubt this
would be a volunteer effort, and may need contractor involvement.  Could
this be combined with the OWASP wiki redesign?


So… the ask from the community, what are your thoughts? Could this be a
viable option to save us time on docs/descriptions and increase quality?
Would there be funds to perform the initial conversion of everything into
markup pages?  Would there be resistance to working in this new model?


Gary
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.owasp.org/pipermail/owasp-leaders/attachments/20160214/71e30da3/attachment.html>


More information about the OWASP-Leaders mailing list