Mailing-List: contact common-dev-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
MIME-Version: 1.0
From: Martin Rosse <mrossesf@gmail.com>
Date: Wed, 27 Jul 2016 12:08:36 -0700
Message-ID: <CAPifw-pW2koJ-o8d1=f5QtijQX1ybPLBF7Mx5-7sbusUAJK0Ow@mail.gmail.com>
Subject: Wiki migration and clean-up
To: Ray Chiang <rchiang@apache.org>
Cc: Andrew Wang <andrew.wang@cloudera.com>,
	"common-dev@hadoop.apache.org" <common-dev@hadoop.apache.org>
Content-Type: multipart/alternative; boundary=94eb2c08905a949bb90538a2be0d
archived-at: Wed, 27 Jul 2016 19:08:49 -0000

--94eb2c08905a949bb90538a2be0d
Content-Type: text/plain; charset=UTF-8

Hi Ray,

The migration is much needed, and thanks for initiating it.

Regarding approaches to cleaning up the Wiki content--my 2 cents is in
favor an approach similar to the Spark cwiki:

https://cwiki.apache.org/confluence/display/SPARK/Wiki+Homepage

My take is that the Hadoop product docs on hadoop.apache.org generally
target (or should target) the audiences you describe in 1-4, while the Wiki
is (should be) primarily for audience #5 or "Hadoop staff"--internal Hadoop
development, product management, QA, etc.

Definitely current Wiki content such as "Overview of Hadoop" and the link
to "Single Node Hadoop Cluster" installation is redundant, unnecessary doc
maintenance, and annoying to come across as a user because you have to
assess its value relative to the same/similar content in the product doc on
hadoop.apache.org.

BTW, I did some random testing of ASF project wikis hosted on
cwiki.apache.org, and the pages for those sites definitely load much, much
faster than ASF wiki pages using MoinMoin. Clearly no surprise.

Best,
Martin


On Wed, Jul 27, 2016 at 10:29 AM, Ray Chiang <rchiang@apache.org> wrote:

> Good to know.  It's certainly easier to set up an alternate location in
> any case and then do a wholesale migration.  It saves from having that
> "under construction" look before it's complete.
>
> I'll get on the appropriate infra@ list and ask about recommendations.
>
> -Ray
>
>
> On 7/26/16 10:49 PM, Andrew Wang wrote:
>
>> Hi Ray, if you're going to do a wiki cleanup, fair warning that I filed
>> this INFRA JIRA about the wiki being terribly slow, and they closed it as
>> WONTFIX:
>>
>> https://issues.apache.org/jira/browse/INFRA-12283
>>
>> So if you'd actually like to undertake a wiki cleanup, we should also
>> consider migrating the content to a wiki that isn't terribly slow.
>>
>> I think cwiki.apache.org is better, but maybe we should ask infra what
>> the
>> preferred option is here. They might be able to help with a content
>> migration too.
>>
>> On Tue, Jul 26, 2016 at 3:27 PM, Ray Chiang <rchiang@apache.org> wrote:
>>
>> Coming in late to an old thread.
>>>
>>> I was looking around at the Hadoop documentation (hadoop.apache.org and
>>> wiki.apache.org/hadoop) and I'd sum up the current state of the
>>> documentation as follows:
>>>
>>> 1. hadoop.apache.org is pretty clearly full of technical information.
>>> My only minor nit here is that the wiki pointer and the Git pointer
>>>     at the top is really tiny.
>>> 2. wiki.apache.org is simultaneously targeted to at least four audiences
>>>      1. Industry Users (broadest sense of Big Data Industry)
>>>      2. Industry Developers (mostly those adding a layer like Hive does
>>>         to MapReduce)
>>>      3. Hadoop Users (those who just want to set up a small cluster)
>>>      4. Hadoop Developers (e.g. using MapReduce APIs)
>>>      5. Hadoop Internal Developers (eventual contributors)
>>>
>>> I'd like to initiate some cleanup of the wiki, but before I even start,
>>> I'd like to see if anyone has constructive suggestions or other
>>> approaches
>>> that would make this transition smoother.
>>>
>>> 1. Some sections, like Industry Users and Industry Developers is
>>>     growing so fast, I'm not sure whether it's worth maintaining in any
>>>     meaningful format. I'd be inclined to make suggestions on where to
>>>     start and let Google take them forward from there.
>>> 2. Organize the developer section based on the pieces a new reader
>>>     wants to learn (new to everything, new to Hadoop, all the tools for
>>>     Hadoop development, "just check out code and go", etc).
>>> 3. Organize the Users section a bit more.  The "Setting up a Hadoop
>>>     Cluster" is grouped well, but I'd perhaps rearrange the ordering a
>>> bit.
>>>
>>> -Ray
>>>
>>

--94eb2c08905a949bb90538a2be0d--