"Current Location?"

[crschmidt]

Current location is cool. I've wanted to upload it myself in various capacities for a long time.

But linking somethign that people will typically type in as free form to Google Maps is stupid. People will type 'Home', and linking that to google maps won't get you anywhere!

So, I went through and fixed my style so that the location would no longer be linked to Google maps, as well as reordering the 'currents' so that location would be on the bottom instead of the top.

So now, when I type 'Home" in my current location, it won't be displayed as "Maps.google.com/?q=Home" -- because it's not. It's home, in Cambridge. Of course, the rest of you will still see the links on your friends pages, but I've never optimized my journal for the users who are reading through friends pages: I've optimized it for *me*, and for my journal view pages.

I can see why this was done, but I think this is a case where the engineers (or engineer *cough* Brad *cough*) should have pulled in their usability person before pushing it live :p It's still a pretty cool little box, but I don't like to see it off to what I see as a negative start. Then again, is there much on LJ that I *don't* see as a negative start? ;)

Updated code is in Entry::print_metadata in my custom layer if anyone wants to crib it.

Comments

[liv]

Thank you so much for this! I have never been able to work out how to customize metadata at all, because Core just says, iterate over some keys, and I can't figure out how it knows what those keys are. Your Entry::print_metadata is really helpful. And you are absolutely right, the link to Google maps sucks, so I've borrowed your code to get rid of it.

Do you know the answer to this query, by any chance? I've done something hacky to change the wording from Location, but presumably there ought to be a proper variable for it.

[rho]

And even if people don't just put in "home" and actually put it in the name of the city, then the link to google maps will still suck. Let's say you'd just put "CAmbridge" as your location. Wikipedia lists 24 different places called Cambridge. Or for me, there are at least 17 different Lancasters (and one person on my friends page who lives in one of the other ones (Lancaster, PA).

This problem is made even worse because I believe that Google Maps displays search results dependant on what country the search is being made from. If I search for Lancaster or Cambridge, it takes me straight to the English versions (and that's Cambridge, Cambridgeshire, mind, not the little village of Cambridge, Gloucestershire). I rather suspect that this won't be the case if you search within the USA. What this means is that even if the link shows up accurately and as a useful link for you, you've still no way of knowing that it's going to work for everyone else.

And it gets worse. If I search Google Maps for "Boston", it quite reasonably assumes I want the major city in Massachusetts rather than the small town in Lincolnshire. What's bad though, is that "Boston, Lincolnshire" gives me a list of places called "Lincolnshire" with Boston, MA. Instead, I have to search for "Boston, England". And if I'm looking for a specific Washington, nothing less than "Washington, Tyne and Wear, England" will get me where I want to go.

This is all sensible behaviour on the part of Google Maps who obviously have to disambiguate somehow, but it makes the link from the "current location" field useless at best and misleading at worst. Very, very few people are going to want to type out three tiers of address in trying to specify their location, or to test how their search performs in different countries.

The other problem I have is that most people are only going to have at most two to three locations (at most) which they regularly update from. Either many people aren't going to use this, or there's going to be a whole lot of repetitive crap being added to entries.

The feature would be most useful for posts from mobile devices, except that these generally don't support currents well. Phone posts can't do currents at all. Posting by email you probably can do this, but it doesn't seem to be documented (yet?) and is something of a faff anyway.

(And the other other problem I have is that the web update page is now formatted so that the "options" section has one column of 6 lines and one column of 4 lines, rather than two of 5 lines as would be sensible.)

[rho]

Oh, and further further gripe. Google Maps only covers about 5 or 6 countries. Search for Moscow or Paris and you get laughable results. The implications of this for a site with a global userbase should be obvious.

[crschmidt.livejournal.com]

Sure, but what solution is better? Google Maps is the only solution that allows *any* kind of external user interaction. I hardly think LiveJournal is in a position where they have the resources neccesary to pay the multi-million dollars TeleAtlas/NavTeq want for data rights ;)

[rho]

To do usability and i18n testing before putting things live :)

Even if we assume that the google map link is useful for people in the US or UK or whatever (which I think we both agree that it isn't, but let's assume that it is) then it almost certainly isn't useful for people within Russia (well, unless they travel outside Russia a whole lot). In fact, it makes the feature less useful because there are radom useless links littering thigs up.

It's not a case of "people from $country are getting more than us" (as are phone posts, for instant) so much as a case of "our usability is being decreased in order to increase usability in $country".

A possible option would be, for instance, having a "link my location too google maps" check box on edit info, and then take its default value based on whether google actually maps the country that person has set. Possibly not the neatest solution, but it would solve the problem as I see it.

(Of course, a better option would be to get rid of the link all together, but I'm writing this comment based on the assumption that it's a good thing.)

[mendel.livejournal.com]

Moscow? Paris?

The idea that an American Internet company shouldn't roll out a feature at all because it doesn't work as well for people in other countries is silly. "If Europe can't use it, no-one can use it" is sour grapes.

"Usability" is like "intuitive" -- it means "I don't like what was done" unless you have a usability study. Otherwise you're just guessing what other people might do, too. My guess is that people won't care that it's mislinked, and linking it will lead people to provide useful information in it so that the link works. But my wild guess is as useful as everyone else's, including 6A's.

That said, the first step of a usability study is figuring out the target audience -- and for LiveJournal, we're not them.

[rho]

Satellite images both. Which isn't what LiveJournal links to. If I actually put Moscow or Paris as the text of my "current location" then I would get a link to a useless page.

And let's be clear here: I'm not saying that if some people can't use a feature then nobody should be able to. I'm saying that if there's a feature that only some people are able to use, it should be designed so it doesn't detract from things for those who can't use it.

Speaking for myself, I would much rather be able to enter plain text or links of my own, rather than having a link forced on me, especially if the link it gave was useless. I obviously haven't done any studies, but I would suspect that in countries not supported by google maps, many people wouldn't care either way, but of those who did care the vast majority would agree with me.

Of course, there are a whole lot of questions that follow on from that. How much does it detract from the feature for those people not covered by google maps? Is the agravation it would cause large enough to be worth doing anything over? Is the aggravation for these people outweighed by the axtra feature available for people in areas that are covered? Does it make better business sense to focus more on the core customer base (the USA) or try to ensure increasing appeal for users in other countries? Would the amount of effort required to program a fix to allow some people to use the feature but others to not have to be worth it?

And so on and so forth. And to these questions, I don't know the answers. It's perfectly possible that with all different considerations factored in, the current solution is the optimal. My wild guess is that it isn't, but that is only a wild guess. However, I do think that there is a genuine issue which shouldn't just be written off as "silly" or "sour grapes".

[mendel.livejournal.com]

Satellite images both. Which isn't what LiveJournal links to.

Stop feeling sorry for all the stupid users who can't figure out how to click on "Satellite image", or that can't understand that a map can't figure out where "Home" is. You figured it out, I figured it out, I bet other people can figure it out as well. Europeans have demonstrated an ability to figure out things like American dollars and how to call American phone-post numbers, they'll figure this one out too.

People who don't like how it works but need to indicate their location can continue indicating their location in exactly the same way they did last week.

[crschmidt.livejournal.com]

The searches from the US for "Cambridge" actually do return "Cambridge, UK" -- I don't think that there's any location-based differences in the searching.

Although your points all apply. I really just think linking this to Google Maps by default was a poorly thought out idea. (This is not a surprise, since it seems as if it was tacked on at the last minute by Brad or someone else who thought "Ooh! Google maps! that's popular!")

In general, I like the "current location" box -- I probably won't use it except when I'm out and about, but that's relatively *often*. Also, note the variety of things that "Current Music" is used for -- everything from "The fan in the window" to a song title. "In the office", "In bed", etc. are all going to be things people will write, and it may actually prove useful in some cases.

But the "Linking to Google Maps" part is broken. However, for LJ features, that's pretty decent: I like the idea, and most of the implementation. That puts it above about 95% of the other features LJ has released. ;)

[mendel.livejournal.com]

Like "usability" and "intuitive", "broken" is just an expression of personal dislike. Does it work the way it was designed? Of course it does. Why not try to put together some thoughts as to why it's worse for the whole userbase than not linking it?

[crschmidt.livejournal.com]

1. With the current setup, there is no (non-hacky) way to use the text of the location on its own. This means that users do not have the ability to easily edit a style to do something interesting with the text.
2. Since the text is automatically turned into a link, users can not link it to something else of their choosing. For example, a role playing game might wish to link to a specific area of a non-earth based map, but this isn't possible, because the text is automatically linked. (Even with striptags, I don't think you'd end up being able to get what you wanted out of it.)
3. The lack of example input means that users will tend to enter content which is not designed to fit the Google Maps link. Similarly to Current Mood and Current Music, it is likely that users will treat the location as free form data, which, when linked to Google Maps, will not provide any useful results (and is much more likely to produce non-useful results, as in the case of "home").
4. Relatively few LiveJournal users are willing to give out information which can uniquely identify them online, much less information that could uniquely identify them in meatspace. The Google map link will prove useful only in the case where a user *is* willing to identify themselves in realspace (assuming that they get past the initial ignorance of how they need to enter data to do so). For any user who is not willing to give out this information, this link will not be useful -- and I would be willing to bet money that there is a much larger percentage of users who would not be willing to than would be willing to. (Hell, I publish my address online, I publish gps logs online, but I don't want to type my address into "Current Location".)

Linking to Google Maps is at the very least stifling the ability of users to creatively include content in the current location field that might otherwise be much more useful due to the inability to include non-text data in said field, and I still maintain that based on my (admittedly non-normal, but I think relatively educated by this point) knowledge of LiveJournal users, more often than not, the link is not going to provide any useful results.

I could be wrong. I often am. I just don't think that I really am, especially with respect to number 4 in the list above, and that is the biggest blocker in my opinion, and the reason I think that the Google Map link is less useful than free text.

[mendel.livejournal.com]

I can see why this was done, but I think this is a case where the engineers (or engineer *cough* Brad *cough*) should have pulled in their usability person before pushing it live :p

What a waste of money that would be! What do you think would happen when they pulled in a "usability person" -- a controlled rollout to 1% or 10% of LiveJournal users selected randomly? For something that has about zero customer impact? No-one's going to stop paying LiveJournal because the "Current Location" field is linked to Google Maps; why not see how all the users use it, and react from there?

Usability studies are cool and have geek cred, but that doesn't mean they're good business decisions, especially for a single trivial feature.

(And remember, much of LiveJournal is just as smart as you are. You figured out that you might want to specify something more specific than "Home" in the location, and I suspect you've also figured out that if someone else writes "Home" it's not worth clicking on. Everyone else is figuring that out right now too.)

[sherm.livejournal.com]

I think by "Usability Person" he means "someone to tell Brad that it doesn't fucking work" rather than "someone who will conduct a usability study."

[crschmidt.livejournal.com]

Right. I figure that's part of the reason that LJ is now owned by 6A: so LJ doesn't have to have all those people just for this site, but rather they can have one UI/Usability person who knows what they're doing to tell people "fix this" before it goes live.

Or it could just be released, with a that'll be fixed afterwards, I suppose. I just never expect that 'broken' things will get fixed, regardless of how broken I think they are -- either because it's LJ (3 years is Real Soon Now) or because it's something that "WORKSFORME".

Whine whine. That's all I ever do.

[sherm.livejournal.com]

It's not just LJ. You should see some of the shit that's running at fortune 100 corporations.

[crschmidt.livejournal.com]

For a long time, I kept wondering when I was going to find a business that actually had a *real* development model: you know, the kind you always imagine exist out there somewhere, with proper development, followed by testing, followed by QA, with documentation, etc.

I kept wondering where the *real* programmers were, the ones who were so much better than me that they write government websites and msn.com and all that jazz.

Then, when I got offered a "Senior Developer" position, I realized -- shit. I am the "real" programmer now, and I *still* haven't found that place. Which probably means that it's just a myth. The people out there who are programming the big iron are just as crappy as I am.

It's a really fucking scary thought.

[(Anonymous)]

This is actually one of your more profound posts. There are lots of programmers out there with poor work habits. Why a company of any size tolerates them is beyond me. In the old days, when a big hard drive was 20 Mbytes and a fast CPU ran at less than 1 Mhz, programmers had to be more disciplined and organized otherwise it was real obvious real fast.

Now with super fast CPUs and unlimited hard drive size, there is no incentive for programmers to write concise code or to be clever with resources. Also management seems hell-bent on end results only. I have heard from one very reliable software development management person at Oracle that over 1/2 of what MSN delivers to the end user is code that has been commented out.

This breakdown in discipline is a global problem. I was talking with a software engineer from HP in San Jose last week and he told me that the programmers in India he knows are the absolute worst in documenting their software. No flowcharts (even after the code is written), no comments, no documentation - they can't be bothered.

"I kept wondering when I was going to find a business that actually had a *real* development model" I suggest you try software development for the medical field. The FDA regulates all medical products brought to market in the US including medical software. They have very specific requirements that include testing, documentation, and quality assurance. You need FDA approval before you can sell it in the US. If the medical company neglects any aspect of the FDA requirements they will be heavily fined or their product pulled from the market After all peoples lives could be at risk. Who wants a pacemaker with faulty firmware.

-v-

[mendel.livejournal.com]

Works great for me. What was the spec?

Usually, crying "usability" means that someone doesn't like something, and to avoid looking selfish by describing it in terms of things they want to do, they explain how some set of fictional users -- usually ridiculously helpless ones -- will be injured, and that's what this looked like.

Chris explained up a bit how he'd prefer it to have been done and why, which is great, but there's a big difference between something that outright does not work and something that one wishes worked differently. Calling the latter the former seriously cuts down on the credibility of the points raised.

[crschmidt.livejournal.com]

I'm sorry, let me be more clear:

I'm selfish, and I think that linking to http://maps.google.com/?q=home is non-ideal because it doesn't fit the way I think of things. I also think that this is true of most people using the site, but that's an obviously-biased opinion. (If everyone agreed with me on everything, my life would be so much easier.)

Hopefully that makes how I feel more clear :)

[mendel.livejournal.com]

I know you're being a bit sarcastic there, but it's not only more clear, but easier to agree with! Let your opinions stand backed up by your own expertise, instead of getting caught up in usability and development process problems which end up amounting to an appeal to authority.

[crschmidt.livejournal.com]

Not entirely sarcastic: I know full well that I don't have the strength of numbers, or evidence, to back me up. I still know that linking to Google Maps *feels* wrong, and it feels wrong in a way that seems like it should be universally wrong for others.

I probably was appealing to authority subconciously because I know that I have no more experience in this kind of thing than anyone else, so I'm no better an authority on the topic, therefore my opinion/expertise in this case is just as easily wrong as I consider others to be. So, if I just whine loud enough, everyone will ignore that and give me what I want, right?

... Right?

[sherm.livejournal.com]

In general, sure.

In this case, blindly dumping the user-supplied data from a freeform textbox into the query string of a service you have no control over is pretty much always going to be stupid, and would indeed be exposed by having someone not in the "gee whiz cool feature!" mindset evaluate it before it went live. You might even say such a person would be concerned with, I don't know, usability?

This kind of rush-to-deploy-nifty-but-not-tought-out stuff has always been a {problem | strength} of LJ's.