[
https://issues.apache.org/jira/browse/SPARK-4992?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14260123#comment-14260123
]
Eric O. LEBIGOT (EOL) edited comment on SPARK-4992 at 12/29/14 2:24 PM:
------------------------------------------------------------------------
Now I see that the problem with Python is more pervasive: the Python
introduction at http://spark.apache.org/docs/latest/quick-start.html does not
respect the style guidelines of PEP 8, for instance (use of camel case for
simple quantities), and shows some unusual convention.
At this point, I think that not shadowing "file", on the main web page, is
important.
Another important shadowing instance is the builtin "max" function, which is
shadowed, in an example at http://spark.apache.org/docs/latest/quick-start.html
(look for "def max"). The function could be called "larger", for instance. I
have not used PySpark yet, but there might be no need to redefine a max
function since there is a builtin (maybe its signature is incompatible with
PySpark?).
More generally, it would be better advertisement for Spark if the Python
examples showed a higher mastery of the language (the errors I see are
beginner's mistakes): they let me wonder about the quality of the Python
bindings. Now, my fears might be just that, and I will still investigate Spark.
I must also say that I am happy that the bindings exist, and I do appreciate
all the volunteer work that went into them and into Spark. :)
was (Author: lebigot):
Now I see that the problem with Python is more pervasive: the Python
introduction at http://spark.apache.org/docs/latest/quick-start.html does not
respect the style guidelines of PEP 8, for instance, and shows some unusual
convention.
At this point, I think that not shadowing "file" is important.
I also noticed that the builtin "max" function is also shadowed by an example
at http://spark.apache.org/docs/latest/quick-start.html (look for "def max").
The function could be called "larger", for instance. I have not used PySpark
yet, but there might be no need to redefine a max function since there is a
builtin (maybe its signature is incompatible with PySpark?).
More generally, it would be better advertisement for Spark if the Python
examples showed a higher mastery of the language (the errors I see are
beginner's mistakes): they let me wonder about the quality of the Python
bindings. Now, my fears might be just that; I am happy that the bindings exist,
and I do appreciate all the volunteer work that went into them and into Spark.
:)
> Prominent Python example has bad, beginner style
> ------------------------------------------------
>
> Key: SPARK-4992
> URL: https://issues.apache.org/jira/browse/SPARK-4992
> Project: Spark
> Issue Type: Improvement
> Components: Documentation
> Reporter: Eric O. LEBIGOT (EOL)
> Priority: Trivial
> Labels: documentation
> Attachments: SPARK-4992.patch
>
> Original Estimate: 5m
> Remaining Estimate: 5m
>
> In the main webpage http://spark.apache.org/, the Python example uses a
> variable named "file": this is a well-known bad practice (because file is the
> name of a built-in class).
> Such a prominently visible example should show a better mastery of Python, if
> Spark is to give confidence in the quality of its Python API.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
