[Liblas-devel] Re: las2txt question

[Etienne Bellemare Racine]

> Etienne,
>
> I am in the process of rewriting las2txt to use the C++ API (so we get the new-style filters, etc).
Great !
> I have everything done except for the header writing portion.  After looking at it a bit, I'm wondering if it is that useful in its current form.  Do you expect that las2txt output of a header should be easily parseable, or you expect it to act much like a comment block at the front of a source code?
I'd say it should be easily readable with other tools like a csv 
reader/editor. That's the reason I personnaly use the las2txt (or 
recommend it to others). For instance load points in PostGIS, or even 
load points in a GIS (when the las2shp is unreadable -e.g. in ArcGIS).
> As it is right now, las2txt's header output does neither very well.  It would be quite simple to have las2txt output libLAS' xml-ification of the header if the user specified a --header argument in the invocation.  Would this be useful?
In my case, I'd say no. It is not usefull because the people that need 
txt file are not going to use an xml parser. I'd say the the design 
should be targetting something simple at least by default. Output one 
file, maybe that option could be useful, for other purposes, but I find 
it quite complex for the name of the tool or the use cases I have in mind.
> Or is the intention that the header in the front of a .txt xyz LAS file is must cursory information that can't actually be used to get back to the original LAS file?
I'm not sure I understand well your question, but I think I see the 
las2txt as an end. I use it because I need a file I can read with a text 
reader. If I want to manipulate the file to create a new las file, I 
should do it using with appropriate tools like python for instance.

Hope it helps you.

Cheers for the rewriting,
Etienne
> Howard
>
>
> On Oct 28, 2010, at 3:44 PM, Etienne Bellemare wrote:
>
>> Good point, I'd like to suggest also the ability to specify the header separated the same way the fields are going to be. It will create a much more consistent file. e.g. for the moment you can't use tab to seperate header even if your fields are. Having the ability to set a name for the header could be a nice feature too.
>>
>> Thanks again,
>> Etienne
>>
>> On Thu, Oct 28, 2010 at 4:22 PM, Howard Butler<hobu.inc at gmail.com>  wrote:
>>
>> On Oct 28, 2010, at 3:16 PM, Randy Bucciarelli wrote:
>>
>>> Hi Howard,
>>>
>>> Is there a way to format resulting precision of values from the output of las2txt?
>>>
>>> Or does 'las2txt' simply output the precision of the values how there were initially stored when creating LAS files?
>> There isn't, but I agree there should be an option to set/override the precision of x/y/z values (independently)  I have rewritten lasinfo and las2las using the new C++ API, and I think las2txt should also be rewritten as well.  In summary output and other venues in those rewritten utilities, I have tried to respect the specified scale/precision of the dimension if it was set in the header.
>>
>> Rewriting las2txt similarly to lasinfo and las2las is in my queue, but I probably won't get to it for a bit.
>>
>> Howard
>>
>> PS, this question should be asked on the list, and I have forwarded it there._______________________________________________
>> Liblas-devel mailing list
>> Liblas-devel at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/liblas-devel
>>
>> _______________________________________________
>> Liblas-devel mailing list
>> Liblas-devel at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/liblas-devel