[Liblas-devel] Point implementation proposal

[Mateusz Loskot]

On 26/07/10 04:09, Howard Butler wrote:
>
> On Jul 25, 2010, at 6:20 PM, Mateusz Loskot wrote:
>> Sounds good to me.
>>
>> A variation of this could be to physically separate
>> responsibilities. By default, I/O operation could be performed on
>> raw data, without any interpretation. Interpretation is left to
>> user. User has liberty to choose to use point interpreter provided
>> by libLAS or to implement his own.
>
> One then asks, what's the point of libLAS then? ;)

In the first case, libLAS reads and parses LAS file
and provides access to its records as they are stored.
In the second case, libLAS extends the former use case
with extra logic of data interpretation and processing.

>> A point interpreter could be passed to reader for better, seamless
>>  integration.
>
> I like the idea of an interpreter being applied to a reader.  My
> schema idea is really a schema (layout) + interpreter.  I hadn't
> described the distinction between the two in my description of the
> schema stuff.

Yes. I see the Schema being part of the interpretation,
but I'm still discussing at higher level.
It just feels to me that interpreter is a user of schema to perform
interpretation.

Best regards,
-- 
Mateusz Loskot, http://mateusz.loskot.net