2007-08-23

More Google Sky

Yes, yes, I know, Google Sky is so yesterday...

Continuing on with the theme of the point and purpose of Google Sky, I see that even Phil Plait has taken the easy route to reviewing it and decided that it isn't as good as some venerable planetarium software. Well dur!

Like I said yesterday, I think it's missing the point to look at Google Sky like that. This is how you should look at Google Sky:

I've spent the afternoon poking around with the internals of the VOEvent broker and the KML documentation and I currently have a live network link (KML) connected to the broker. This means that any OGLE, Robonet-1.0, ESSENCE, SDSS, GCN or other event messages that flow across the backbone will be automatically published to Google Sky.
See where this is heading? See that it's about data and information sharing? See how it's not about trying to be yet another bloody planetarium?

The two big problems I see with the whole Google Sky approach at the moment are this: First, there's no online version. By that I mean there's no sky.google.com in the same way we've got maps.google.com. I think that's a serious downer. Providing a way of embedding maps into other sites is seriously useful and I think it's one of the really big wins of the whole Google Maps/Earth arena.

The second big problem I see at the moment is that, with Google Sky, the only extension service that appears to be available is via KML. What I've not yet seen in the documentation is any hint of any ability to "search" and find things as part of some presentation. If that doesn't make sense, consider an example:

Suppose I wanted to knock up a layer that presents my observing logs. As it happens, I don't record the RA/Dec of the objects I observe, ever. What I do record are names and IDs of such objects. So, for example, I know all the times that I've observed M13. To create a KML for Google Sky I need a way of converting "M13" into RA/Dec. From what I've seen of the documentation so far there's no obvious way of doing this.

Over in the Google Maps/Earth arena, this problem is covered. There's ways of, for example, turning a postal code into a lat/lon. So, in the Google universe, is there or will there be an API for turning Messier IDs or NGCs into positions in the sky? And, if not, why not?

All of which brings me to the point that does disappoint me about Google Sky. There's a project out that that has been addressing the above issues and has been doing it for some time. I first noticed it late last year. It's Sky-Map.org.

Now, at this point, I was going to point to a couple of examples of how you can do handy little things with the Sky-Map.org API, how it has the astronomical equivalent of the post code lookup facilities I was talking about above. Problem is, while Sky-Map.org itself seems to be up and running, the documentation for the API appears to have gone AWOL.

That aside though, let me give one example of how I do currently use Sky-Map and how I use it in a way that, from what I've seen of the documentation so far, you can't use Google Sky. In the markup I use for my observing logs I have a way of marking up a star. When I generate XHTML from that XML I simply turn that markup into a URL that takes the reader to Sky-Map.org, to the location of that star, and I don't even need to worry about the location of the star because Sky-Map.org does it for me. I only need use:

http://www.sky-map.org/?object=<star name>

For example:

http://www.sky-map.org/?object=TYC5575-473-1

This is the sort of stuff I need to make Google Sky useful. I don't care how badly it isn't Stellarium, Stellarium does a pretty good job of being Stellarium.

I want it to be like Sky-Map.org, with bells on, and other things I didn't know I needed.

18 comments:

Paul Sutherland said...

It is good to read a post like yours, Dave, that sees the point of Google Sky and, more importantly, the potential. I can imagine this developing quite quickly into a very useful application that professional astronomers use to help identify objects in their own presentations. Perhaps Google ought to buy up the talent at Sky-Map to help develop it.
I do think Google might have consulted more with amateur astronomers in creating the application. It does seem incredible that it cannot find something as obvious as the Pleiades, whether you call them by that name or by their Messier number.

Dave Pearson said...

But they are consulting with amateur astronomers. This release of Google Earth is a Beta release. That's one of the reasons software developers and companies release Beta versions of their software: to let people who are interested in it iron out such problems.

Paul Sutherland said...

I'm suggesting that there could have been a little more consultation with amateur astronomers before the Beta release. It might be prejudice on my part, but I can imagine a Hubble astronomer failing to spot the Pleiades where an astronomer might have spotted the problem straight away.

Dave Pearson said...

Obviously, but I'm willing to bet that it's just a minor omission from one of the data feeds or something.

Moreover, how much (or little) consultation with "amateur astronomers" was there before release of the beta? Is it just the assumption that a specific issue that's obvious to you is evidence that there was none at all?

Also, what's so special about pre-beta in this case?

I think it's too easy (and a little lazy) to characterise this as a pro vs am astronomer issue. But, then again, I get totally fed up with that daft argument anyway so it's probably best to ignore me on that point.

James said...

As I said in the previous post, lots of potential but I'd rather not learn how to code to get something useful out of this.

And I'm glad Paul couldn't find M45 either, I thought I was going crazy.

Dave Pearson said...

I don't think anyone expects non-coders (or coders who can't be bothered) to code to get anything useful out of it. But I do expect that non-coders will get some useful stuff out of it once coders who care and are interested get to work on making interesting things available.

As I pointed to in this article, within 24 hours of release we had at least 1 interesting addition.

It's not that I think you personally, James, should be interested in it, or excited about it, or whatever. But I do think that quite a lot of people have been looking at this software all wrong (even more so given that it's beta). I do see problems with it, but I think the least-worst problem is "it's not Stellarium" and, yet, quite a few people seem to be making this exact complaint.

Astrogeek said...

Hi Dave;

I think you are perhaps cherry-picking the negative comments about Google Sky. Both my original post and Phil Plait's post about GS indicated that it had its good points; which it does. I sincerely hope that they improve the imagery blending, and I think that it certainly has a place in teaching people about the wonders of the sky, and in sharing imagery and observations, but at the same time, I don't foresee myself coding KML anytime soon.

I'm all for Google providing yet another free tool for online collaboration, and if other people can make additional layers that provide useful features, I'm all for that too.

Dave Pearson said...

I hope I'm not. At least, I know that isn't my motive.

Bugs and oddness aside (bugs and oddness are to be expected, right?) the common theme I've been seeing is "it's not very good as a planetarium programme", I even nodded in that direction myself. I don't think that identifying a common theme, and one that I suspect is missing the point, is "cherry picking".

Or, put another way, Phil (to pick one example) is "scratching [his] head over why they left out so many obvious and necessary features". I'm not because my (possibly very wrong) impression is that that isn't the primary point of the Sky mode for Google Earth. Doubtless it'll evolve that way, perhaps even thanks to other people's work, but I'm not scratching my head because I get the impression that competing with planetarium software isn't the point.

It seems to me that competing with something like sky-map.org is the point (and, on this point, I think it is somewhat lacking).

Astrogeek said...


Pamela Gay likes it
and pretty much for the reasons I anticipated in my comments. It's a useful tool for teaching basic astronomy concepts, and practicing (with the presence of a broadband internet connection) what the students will be seeing in the sky when they get 'out there'. I typically don't drag my computer out unless I'm doing a presentation, and even then I'm usually only showing screenshots from Stellarium embedded in a powerpoint. Most of my teaching is done with a GLP or behind the EP of the Challenger.

Dr. Gay is doing different things than I am, and she thinks it will be very useful for her.

I really thing GS has (or will have) a place in teaching astronomy concepts, and will probably also have an important role in online collaboration and sharing, but I have to agree with Phil that they seem to have left out some important features. Hopefully that's just a function of the 'Beta' status.

Philip said...

Dave-- As I pointed out in my article and in the comments, the Sky layer for Google Earth will come into its own as people write APIs for it and word spreads. It has a lot of potential... but I stand by my opinion that in this first release, that potential is barely visible.

Of *course* it will be compared to planetarium software. The fact is, almost all the astronomy bloggers have made that comparison. It *is* planetarium software, but at the moment lacks the basic features it needs. They should have had those in at the minimum; I think that the average user will get bored with it pretty rapidly unless more features are added, and quickly.

Dave Pearson said...

Hmm, see, I see it more as mapping software than I do planetarium software (and, yes, I know, "here be dragons", there's a real danger of getting into a debate about what those terms mean, just run with me on that score). What I see here is a platform on which neat little tools and hacks can be deployed.

Like I said, I think it makes far more sense to compare it to the likes of sky-map.org than it does to Starry Night and friends.

I guess what I'm saying (and I'm fully aware of the fact that I'm saying this as a jobbing software developer, not as your "average" user) is, far from the potential being "barely visible", I think it's the most visible aspect of it right now.

Al. said...

You asked for a way of converting "M13" into RA/Dec, but of course Google Sky already does this, try typing "M13"" into the search box.

Now I'm only guessing but my bet is that Google Sky, like all the rest of astronomy, is piggy backing this capability off the back of the SIMBAD name resolver service at CDS.

Astrogeek said...

@Al

To be fair, what he asked for was a way of doing that within KML similar to how he does it with XML or XHTML and sky-map.org.

Dave Pearson said...

Well, yeah, it's obvious that it does it inside somewhere, I was thinking more about Google publishing some sort of API so it could be done in code in the same way you can do that sort of thing with postal codes.

Thanks for the SIMBAD resolver link, I don't think I'd seen this before.

Hmm, that does know about M45 so I'm thinking that there's more to GS's search box than using SIMBAD (thinking of Paul's quirk here).

Al. said...

Why does Google have to publish an API, why can't you use the CDS SIMBAD name resolver web services to do it? If you want to do it from code, you can just use SIMBAD directly, most of my stuff does. In fact if you're a Perl person you can use my modules to do it...

Isn't that what mashups are all about after all?

Stuart said...

I echo the point about using SIMBAD to resolve names.

I would guess that Google wanted the astronomers to sign non-disclosure/embargo agreements before launch to keep the project under wraps; I certainly hadn't heard any rumours about it. This may have made it difficult to approach amateur astronomers before launch.

Dave, when I get chance (perhaps over the weekend while travelling) I'll make a kml file for our telescopes. I don't have Google Earth (and don't want to install it) so once I've made it it would be great if you could test/debug it for me.

I suspect (although I have no evidence) that Google may eventually tie-in with data from the Las Cumbres Observatory Global Telescope Network (LCOGTN) which recently bought the Faulkes Telescope project. I have two reasons for thinking this. Firstly, it would be great promotion for the LCOGTN. Secondly, and possibly more importantly, the guy financing LCOGTN made his money at Google so knows the Google folks personally.

Dave Pearson said...

Sure Stuart, I'd be more than happy to try it out.

Dave Pearson said...

Al asked: "Why does Google have to publish an API".

It doesn't have to and, now that I'm aware of that service, that's one less problem from where I'm sat. I could have (and probably would have) used the sky-maps.org service anyway (perhaps that uses the same thing too?) to do the "mash up" thing.

Thing is, I doubt I'm alone in seeing what Sky offers and thinking about how I could use it and then wondering how a certain thing is done and, by extension, looking at what I can do with Google Maps and the related APIs and looking for something similar in the Sky realm. It seems fair and obvious to wonder how you'd do these things and wonder why they apparently don't exist in the "Google universe". Now that I'm aware of an existing service that'll handle it that's personally a problem solved (and thanks for brining it to my attention).