Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Two usability issues in Designation Search
#1
Hi Greg,

Two issues I just noticed with 4.0d:

1. The search returns unexpected results. For instance, a search for NGC 779 returns a few objects containing that string (fine) but also NGC 470, NGC 936, NGC 1087, NGC 2701 and a few more. I see all contain at least one of the digits from my search object but the fuzzy logic by which they are selected is not clear to me. It doesn't appear to be the Levenshtein distance (NGC 679 would have been listed in that case) nor do all the NGC objects having a 7 or a 9 in the name. I'm not sure about the fuzzy logic used, but anyway, a suggestion: if there are results equal to or containing the search string as a substring, then do not do a fuzzy search, it's unlikely they are better matches. I think fuzzy matching is a difficult task on catalog designations: NGC1233 and  NGC1234 are only a character apart but still very different objects. Or, if other users prefer it, perhaps have a checkbox to enable/disable fuzzy searching.

2. Following a successful search, another one that returns no results does not clear the results from the previous one. For instance, search for NGC 779 (which will return some results) then for Abel 779 (a real object not in the ST4 database) (EDIT: I should have searched for ACO 779 instead)  for Canada, for that matter. The results from the previous search are still displayed and there's no indication the search was unsuccessful in fact.

Thanks,
Razvan
Reply
#2
If you are looking for the galaxy cluster it is Abell 779 but should be searched for under ACO 779 which is the current accepted designation for these objects. The Abell designation is only for the planetary nebulae.

Owen
Reply
#3
Hi Owen,

That's correct, I learnt that after I made the post seeing that other Abell clusters were not found (I was pasting designations from other places) and it looked odd. I'll amend the post but it the real issue with the search results not cleared stays.

Thanks.
Reply
#4
Hello,

The search engine normally works the way you would expect, not doing a fuzzy search on something definitive like an NGC number. But there are other factors at play. If you open the other results in the Object Info you will find that they are not "normal" NGC objects. The first few are supplemental database objects. A supplemental database object can be described by any string, so they are always matched via a fuzzy search. I did not intend for supplemental database objects to have NGC numbers, but some have crept in for various reasons, usually via 3rd party target lists. Others are Quasars associated with NGC numbers in an alternative designation, which are normally strings so they are fuzzy searched as well.

The bottom line is that there are many many designations beyond the NGC, for everything from dark nebulae to odd discovery designations for open clusters. It is not a simple matter to have a global search for all of these possibilities. Intuitively we want NGC xxxx to be definitive, but other possibilities intrude. So the solution is to always put the definitive result at the top. It can often be helpful to also display these other results.

When people enter a search string, such as Canada, there is the possibility that they misspelled it. So in the rare situation where nothing matches at all, the original search entry is highlighted in the search area so that the user knows what they typed (otherwise it will disappear). By highlighting it, it is easily typed over or edited. I agree that the results should be cleared in this case and I will add that.

BTW for an NGC number you can enter just the number, e.g. "7741" and it will assume it is an NGC.
Clear skies,
Greg

SkyTools Developer
Reply
#5
Thanks, Greg (and just to be sure it's clear to anyone who reads the thread, I used "Canada" as a search string just as a string for which there's no result, not that I'd search for it Smile
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)