juneau-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From James Bognar <jamesbog...@apache.org>
Subject Re: New blog post: Improvements made to the internals of the Serializer/Parser APIs
Date Sun, 24 Dec 2017 20:18:54 GMT
Hi Gary,

Good question.  Story time.

Juneau has been around since 2007 as a cleanroom internal IBM
project.  Jackson wasn't around back then.  For many years, it was the
legal alternative to Jackson for IBM projects due to Jackson having a "do
no evil" licensing clause which conflicted with the IBM ethos (ha!).

Anyway....Juneau and Jackson have some similarities, but Juneau was always
more of a generalized serialization framework geared towards multiple
languages, whereas Jackson was more focused on JSON.  As a small example,
Juneau has ObjectMap/ObjectList instead of JSONObject/JSONArray.

IBM did scalability comparison testing between the two projects in 2014.
What they found was that Juneau outperformed Jackson by 20% on small data,
and performed even better as the scale increased.

On Sun, Dec 24, 2017 at 2:48 PM, Gary Gregory <garydgregory@gmail.com>

> Hi,
> Why did Juneau roll its own serialization framework instead of using
> Jackson (for example)?
> Gary
> On Sun, Dec 24, 2017 at 11:02 AM, James Bognar <jamesbognar@apache.org>
> wrote:
> > https://cwiki.apache.org/confluence/pages/viewpage.
> action?pageId=74691612
> >

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message