[Cluster-tech] spider's DE length

[Felipe Ceglia - PY1NB]

Hi Dirk,

I understand your rant. The whole thing has grown, and almost fell out 
of control. I will suggest N1MM and others to change the way they 
identify these robot spots.

Those robots spots are not meant to leak into the dx cluster network. 
That would be a nightmare. That is why we are looking for a safe way to 
isolate those spots from the rest of the dx cluster network.

73,

Felipe


Dirk Koopman G1TLH wrote:
> Felipe
> 
> This is where I have a little rant, I apologise in advance.
> 
> <rant>
> I am stuck between a rock and a hard place here. The format of the USER 
> presentation is fixed - period, unless something radical in bloody 
> stupid spot scraping software has changed that is, I am afraid, how it 
> is going to stay.
> 
> You have NO IDEA how much grief I got when I first started out on this 
> (over 10 years ago) because the USER presentation was not *exactly* the 
> same as AK1A.
> 
> -# is, as far as I am concerned, invalid. I suspect that, for many 
> logging programs, that is also true. I can tell you now that if I 
> allowed the field to expand (an obvious and very easy mod), several 
> logging programs would fail, because they expect the data to appear in 
> *fixed* character positions.
> 
> This is because the authors either don't have a clue or (perhaps) the 
> tools to parse the data in a more flexible way. There are actually 
> fairly simple regexes that could get the data out from this crummy 
> format (for computers), but a lot of authors don't either know or 
> understand about this.
> </rant>
> 
> I fundamentally disagree that -# is useful or desirable. I will filter 
> it out in future releases. If N1MM et al want to distinguish between a 
> robot spot then they can do that, very easily, by looking at the comment 
> section. Can I also point out that the network is already full of spot 
> suckers, web clusters and other specialist robots and they do not cause 
> this sort of problem.
> 
> Has *anyone* considered the consequences of several people running this 
> software - all over the world - during CQWW (or any other large contest)?
> 
> As I understand it, whilst DXSpider seems to cope OK, other dxcluster 
> node software is (already) having quite a bit of difficulty keeping up. 
> Also it potentially causes a quantum leap in the level of protocol 
> traffic - which sysops will have to pay for.
> 
> YMMV
> 
> 73 Dirk G1TLH
> 
> Felipe Ceglia - PY1NB wrote:
>> Hi Dirk,
>>
>> First, sorry, the DE info is being cropped at 7 chars...
>>
>> The actual scenario is that we are feeding an isolate node, 
>> k4td.com:7300, with spots coming from the reverse beacon network.
>>
>> Those spots come tagged as callsign-# so, for example, VE2WU-#, being 
>> 7 char in length is cropped as: VE2WU-# (the ":" is missing, see 
>> below). Some logging programs are rejecting these spots, most likely 
>> because the missing ":" .
>>
>> DX de K4TD-#:    14044.9  K2LP    RBN 23 dB  25 WPM  CQ <K4TD-0  1245Z
>> DX de N4ZR-#:    14045.0  K2LP    RBN 15 dB  22 WPM  CQ <K4TD-0  1246Z
>> DX de VE2WU-#    14027.4  K2ETS   RBN 16 dB  20 WPM  CQ <K4TD-0  1247Z
>>
>> I am looking for a way to fit the string on it, keeping the #.
>> It would be important to keep the #, because it lets software like 
>> N1MM tag spots which have been generated by robots.
>>
>> 73 TU,
>>
>> Felipe
>>
>> Dirk Koopman G1TLH wrote:
>>> Felipe Ceglia - PY1NB wrote:
>>>> Hello folks,
>>>>
>>>> Our spider cluster is cropping the DE callsigns on spots to a 
>>>> maximum of
>>>> 6 chars. Is there a way to raise that limit, or should the code be
>>>> tweaked directly?
>>>>
>>>
>>> Please can you give some examples from the debug file. I would like 
>>> an example of what comes in and the resultant output.
>>>
>>> This is a limit for certain operations (dupe checking and such) but 
>>> the actual spotted called length is 20 characters (from memory).
>>>
>>> There are restrictions on callsign, but they are to do with format, 
>>> rather than length.
>>>
>>> Dirk G1TLH
>>>
>>
> 
> 
> _______________________________________________
> Cluster-tech mailing list
> Cluster-tech at dxcluster.org
> http://mailman.tobit.co.uk/mailman/listinfo/cluster-tech
> 

-- 
Felipe Ceglia - PY1NB
-------------------------------------------------------------------------------------
PR1T team member /// Rio DX Group member /// Araucaria DX Group member
http://www.dxwatch.com /// http://reversebeacon.net /// 
http://riodxgroup.dxwatch.com