I think that changing them for the first release makes sense otherwise you
will need to decide on deprecation or backward compatibility, etc.
None of which is a big deal but starting for a first Apache release with
the names set seems nice and clean.
That said, a lot of podlings do come in and con
I don't have a strong opinion about the proposed renaming but for
end-users and consumers it may be easier to not change too many things
at once. Some renaming (like packaging and OneTableXX) is unavoidable
to be compliant with ASF branding and release guidelines and it is
good to do them early but
Thank you, Tim and Larry!
I appreciate the suggestion and the feedback. I’m on board with using
Conversion* as the class names.
Best,
Ashvin
On Fri, Mar 29, 2024 at 8:24 AM larry mccay wrote:
> I like the proposed renaming away from *Client's.
> SourceReader and TargetWriter seem to imply that
I like the proposed renaming away from *Client's.
SourceReader and TargetWriter seem to imply that there are SourceWriter and
TargetReader though and otherwise it is kind of redundant.
At the same time, Source implies reading and Target implies writing, I
think.
Given the ConversionController sugg
Hi Ashvin,
I like your recommendation and reasoning for not including metadata in the
name. I think SourceReader and TargetWriter are a good option. Including
'table' may make the names too long when also prefixing with a format name.
I agree on the need to change OneTableClient as well. Conversi
Hello,
As we’re considering the renaming of “clients,” I’d like to bring up the
OneTableClient as well. Given its role in overseeing the lifecycle of a
conversion request, it seems fitting to rename it in line with the source
and target clients. My suggestion would be to opt for either
ConversionC
Hi Tim,
The confusion regarding HTTP clients seems to stem from the term “Client”
rather than “Source” and “Target”. Therefore, rather than substituting
“Source” and “Target” with “Metadata”, would it be better to replace
“Client” with "Reader" and "Writer" or "TableReader" and "TableWriter"? The
Hi Everyone,
As we are working through the renaming, I would like to suggest that we
rename the SourceClient and TargetClient interfaces to MetadataReader and
MetadataWriter respectively. I think that some users and developers may
think of these as http clients.
What does everyone think? Any othe
Hi Jesus, Tim,
Thanks for the feedback.
Tim, the ‘Internal’ prefix sounds good to me, it was my initial thought as
well. It is simple and descriptive, and decouples the library’s name from
its functionality.
My only worry is that it might lead to lengthy class names. However, I’m
willing to test i
One other option is to prefix with "Internal" or something similar instead
of the XT so it's clear it is our intermediate representation.
-Tim
On Fri, Mar 22, 2024 at 5:29 PM Jesus Camacho Rodriguez
wrote:
> Thanks for starting this discussion, Ashvin!
>
> I think the proposal makes sense. Othe
Thanks for starting this discussion, Ashvin!
I think the proposal makes sense. Otherwise, we may find ourselves needing
to explicitly reference the classes using the namespace too often for
common names across table formats.
-Jesús
On Fri, Mar 22, 2024 at 12:09 PM Ashvin A wrote:
> Hello All,
11 matches
Mail list logo
