I especially love this part:
I've been a happy Google
customerproduct for a long time, because Google tools used to enhance the Internet. But as Google ships "the Google part" of its new Google+ identity, it's breaking the Web it once helped build. I can't take it anymore.
[...] This might be the Googlest thing I've ever seen. First of all, he designates himself a "user/reader," which is just spectacular. And he goes on to say that he'd rather see a Google+ post about the article than the article itself, because, and I quote, "I can comment here without having to jump through hoops." Hoops like reading the article?
Google is great for work. My work personality is much more measured, flat and bland than my real self, and Google tools are great for expressing that.
There's a lot of well-deserved G+ hate in the above-linked article, but the "reposts are more important than the original" part really riles me up.
This is a pervasive, toxic notion that has metastasized in the interwebs in the last few years: that the most important thing to know, when you see a thousand-times-reposted link to something, is the person you most immediately saw the link from. The second most important thing is the first person to repost that link on whatever service you are logged in to at the moment. A far distant 99th priority is the person who actually created the thing in the first place.
One DJ says to another, "Hey, want to go see a movie?" The other DJ says, "I dunno, who's the projectionist?"
This is the kind of behavior that Google+ is encouraging now, and of course they didn't even invent that, they just ripped off this model from Tumblr who seem to have pioneered it by showing the entire thread of attributions by default, and emphasizing the first and last -- but stopping cold at the walls of the Tumblr garden. To link to an actual creator, you have to take an extra step, so nobody bothers.
The reason is obvious: these sites want to keep you, the product, inside the walled garden. And hey, that worked amazingly well for AOL in the 90s, when they learned how to "engage" the Internet. Oh wait, no it didn't.
I'm sure I'd be even more pissed about how awful G+ is if I actually used it, but I only log in about once a month, because it's a ghost-town. The people I actually communicate with on a regular basis all have G+ accounts but almost exclusively use Facebook instead. (The exception being my friends who are actual Google employees, and I'm sure the electrified genital cuffs have something to do with that.) Facebook is a horror in its own innumerable ways, of course, but it's still less
user- product-hostile than G+.
JWZ - you don't need to be told this, but for the benefit of any other readers with less knowledge of the ways of Google, here's something I just posted on by own FB, as my own response to Jon's article:
How to make Google results usable again:
1. Go to http://www.google.com/reviews/t
2. Under "Manually block a site", type in "http://plus.google.com/".
3. Click Block Site.
Now, when you search for a page about X, the first result will be... a page about X. Previously, the first two or three or six dozen results were Google+ posts linking to the page about X and lots of comments by idiots you don't know who wouldn't know X if it bit them on the arse. The improvement is immediate and beautiful.
It's nice that that works, but it requires that you be logged in to Google while searching and/or allow them to store cookies, which is unacceptable to me.
Good point. I find the usefulness worth the soulrape, but I can understand why this is not a universally acceptable trade-off.
Alternatively, you can manually add that as a search parameter, or edit your browser's search provider settings or use a greasemonkey script to add an extra 'q' element to the form.
document.getElementById('tsf').innerHTML += "<input name='q' value='-site:http://plus.google.com/'>"
Do people actually use Tumblr for anything where that social detriment actually exists?
The only people I know who use it are using it as a sort of Livemygeocitiesjournalspace for porn and things esoteric enough that they might as well be porn, where I guess if you actually follow it daily you might care to know which friend of yours found Particular_Insane_Thing.
But even then, I think attribution or original-source-link usually makes it into the body of the posts.
I use tumblr as a porn filter (esoteric or otherwise), and at a certain level I could care less about the attribution of those images. That said, I try to keep all the links that come with a reblogged picture, occasionally add a link if I've been interested enough in a picture to track down a source (not necessarily the source), and I will link to the site if I bring a picture in from outside tumblr.
To be fair, tumblr has gotten a lot better, in that attribution is now the default, and the user has to take extra steps to remove it. (Or configure their blog theme to not show it to non-logged-in users.)
It is really annoying sometimes when you want to find more of that erotica/esoterica and there's no trail to lead you.
By which I assume you mean "The attribution of the first person to copy it to Tumblr is now the default", to which I reply, "Who gives a fuck?"
I posted about this on Metafilter a while back, and the part that really infuriates me is that they strip the filename from photos you upload so even that little metadata clue doesn't exist to facilitate working around lazy/unresponsive reposters (I used to try using the Ask function to get some info from the original poster on the source of images I wanted to further research, but gave up as hardly anyone bothers to respond and those who do just say 'found it on the internet somewhere'). I think it's pretty clear that Tumblr deliberately works to sever the link between content in their ecosystem and its sources on the larger web and they deserve to be called out on that.
At one point I think you could preserve the filename by using the 'URL' source to hotlink a photo on another (i.e. your own) server rather than uploading it, but I believe now even those are uploaded to Tumblr HQ and assigned a new filename.
Bullshit. Every post on tumblr has a source field. If it gets filled in, it gets carried around for the life of the post. If someone deletes the attribution from the post body, tumblr silently reinserts an attribution for you. Posting using one of their automated means (bookmarklet, etc) automatically populates the source field.
If someone is posting data already divorced from its source, say, uploads an image from their harddrive or manually copies and pastes some text, and doesn't provide an attribution, then yeah, you're out of luck. But what more can software do at that point? They've already provided an easy way to do the right thing.
Sure, they probably should preserve the filename, but what are the odds that will provide any useful data? (That pasting the image into google image search won't already give you?) The filename was probably already CMS-generated -- the vast majority of pretty-photo blogs are fueled by flickr explore or artist sites driven by drop-in image gallery scripts.
(And while you're talking metadata, I don't suppose you've tried checking the EXIF for clues?)
Right, but this is still depending on the user to provide source in good-faith (unless using the bookmarklet, but I question what percentage of users are doing that; I didn't even know it existed until you mentioned it and it's not exactly easy to find on the site), and only covers the subset of posts which are actually made at the time-of-discovery, and not made from a file saved locally and uploaded later. While I was surprised to discover just now that Tumblr does preserve EXIF data, that's both not consistently used by creators and not well-known or accesible to the general public. My main point here is that they could be doing this in a way that by-default respects the wishes of the initial creator/namer of the image, rather than depending on the re-poster to do the right thing, regardless of how easy it's made (because a quick stroll around the internet will reveal utterly profound laziness and indifference).
All of this could be moot if they just preserved the file name (or appended their unique id to the end), which is the one reliable, well-known and accesible form of metadata for images. I imagine you're aware how deficient a tool google search-by-image is compared to an actual web search, and having an author name or work title to start with gives you a much greater chance of success in tracking down the original source of an image, or at least some more information.
You're right that countless other CMS' are guilty of this destruction of metadata, but Flickr at least is a different case both because its files are generally uploaded by their creators, and their filenames contain a unique string that can trace you back to the account, though of course that's non-obvious.
"My main point here is that they could be doing this in a way that by-default respects the wishes of the initial creator/namer of the image"
Wait, what way is that?
I mean, if people aren't using the bookmarklet, that sucks because tumblr-ing any other way is doing it the hard way.
But it seems like your argument hinges on people naming their files something human-identifiable, such as artst_name-work_title.jpg, and how often does that actually occur? The creators aren't doing it -- I never get files from photographers that look like that. And anyone who can't be bothered to attribute on their blog isn't going to do it when they originally rip it off the web.
It's true that nobody ever gives their image files sensible names, but it's also true that searching for 5234534987695.jpg or img_4725.jpg often works.
Well I think you're being obtuse, but for my last response, to answer your question, how often does that actually occur : fairly often. You don't think people on the internet use filenames to provide information about images? A few samples GIS searches will reveal that the great majority of returned files have something useful in their name, and in face are probably discoverable in search largely for that reason. No, it's certainly not a universal behavior but my point is that the designers of these infrastructures could enable the option of creators including some simple, useful, widely-accesible metadata in the form of a filename which they could be somewhat confident will propagate along with their work; as it is they are destroying that possibility (unless you think it's justifiable for them to live in a fantasy world where everyone knows what IPTC data is and how to access it).
Of course this all YMMV territory, and probably depends on the circles you traffic in. My assumption was that if Google has indexed an image filename, it has probably fed that file into their magical image-recognition databanks? I've frequently tried searching on filenames; it seems to stump the search engine about as frequently as "search by images."
Anyway, I agree preserving filenames would be nice and we should organize an write-in campaign to firstname.lastname@example.org (whom I've found to be very responsive) immediately.
No, the actual source. That changed about two years ago I think.
Oh come on, yes: tumblr is a disaster, a train wreck of people tracking down 4chan derived stuff + weird porn.
...and that kind of bad habits, may even fit the bill.
Original content? there is little, mostly GIFs out of movies. Some graphic artist.
Creative people actually like to use it this way: they drop content in the stream and wait to see how many people likes or reacts to it. The consumer doesn't really bother who's the first in the chain. but the first in the chain GETS NOTICES.
It's kind of like asking the zeitgeist if you're on the right track or not.
(Spambots may pretend they're part of the zeitgeist, then)
So the real deal is again and again, people over g+ always defending it like it's some shiiiiiiiiiny rrrrrring.
And pretending it's something.. mhh... serious: some sort of linkedinbooktumblr. Now... THAT's a train wreck.
Then again tumblr COULD be a serious platform for blogging in the era of blogger.com demise, (seriously, google is killing it, very slowly and extremely painfully, death by feature update starvation + captca extravaganza) but it doesn't intend to, and resists those who use it as such.
TRICK: on tumblr I use Right Click → "Search Google with this image" a lot.
( Check it out on http://www.google.com/insidesearch/searchbyimage.html on the last panel, bottom right, there are chrome and firefox extensions to install )
Actually, the reason that I care more about who is doing the reposting is that I only have time to follow a limited number of links. Knowing who reposted helps me to estimate why that person thought it worth reposting, which helps me judge whether to invest the time to follow the link.
It is only when I have decided to follow the link that the original author becomes important to me.
This is similar to reading your mail. Everyone has an "uncle Fred" who seems to have unlimited time for forwards and "you gotta see this site" messages. If you're like me, you are a lot less likely to visit the site when the message came from uncle Fred.
That being said, I have noticed that GPlus and Facebook appear to be trying to recreate the old AOL experience.
This sort of thing is why I *hate* Tumblr and the need to follow a link back through a thousand steps to the source. Twitter, on the other hand, has gotten it partially correct with the built-in-retweets. It would be a hundred times better if they flat out banned the "RT" type retweets though. But that's okay, I just tend to block people that use them too much.
I prefer RT style retweets, because they usually display the original link and not the t.co interceptor that built-in retweets display. Which is another reason why I want to know who retweeted something more than who originally tweeted it. With multiple levels of url shorteners, you may not know where a link goes. I only want to follow links that are somehow endorsed by someone I trust as being reliable.
Everything gets the t.co runaround these days. They're very good at pushing anything that even looks like it might be a link through that. (Malformed http://? Sure. foo.com? Yep.)
Of note, further down in the article:
I think he's laboring under a misconception here. It's shipped. Where by "it" I mean they have shipped the announcement that they are eventually, really, going to do something about RealNames, which by the three-to-ten-steps game of Telephone most people get their news from means that a disturbing number of people now think that G+ allows pseudonyms even though it doesn't and is showing no signs of doing so. Mission accomplished! (Myself, I'm looking forward to version 2.0 of the announcement, which will presumably have even more features.)
The points about attribution are valid and very important.
Apparently you and Jon Mitchell aren't using G+ the right way.
People who do it the right way put hundreds of people in their circlesto get the most content possible. I don't do it the right way either.
I can't help but feel that Jon is deliberately misunderstanding the "jumping through hoops" comment. The user/reader wasn't talking about reading, he was talking about posting. Here are the hoops one has to go through to post a comment on an orignal article after discovering it on G+:
o find a way to login (do they have openID? must I create an account?)
o learn the BBS code
o come back and look for responses
Anyway, what's wrong with calling himself a "user/reader"? He's not calling himself a customer. Is it really an excuse to go off on a "you're not the customer, you're the product" rant?
My goal for G+ is to engage with mainstream news organizations about things they're missing, and that's been going okay. A few major news TV programs ran some things that they might have missed, possibly because I told them about them in the context of this new gee-whiz social media latest thing. However, I've had more failures than success so far. We'll see in another six months or so whether it's actually worth it. A lot of news orgs dipped their toes in and haven't been heard from since November.
As for the obvious walled garden, missing attribution, elided source mistakes, well, those are going to blow up sooner or later. I predict much "pivoting" will occur after the users start making fun of the stats.
Tumblr - isn't that like blogging for people with nothing to say?