The new extension from Performancing.com for posting blog articles seems to be a super slick system.  I’ll be testing it out and will update later with some more thoughts on it.

Performancing for Firefox | Performancing.com

In addition to the points I brought up in my previous post on Riya, I’d like to see:

Tight Picasa Integration
I want the Riya Uploader to automatically upload files detected by Picasa’s Media Detector. I also want to be able to pull the metadata that Riya generates back into Picasa.

Thumbnail Chooser
I’d like to be able to choose which photo to use as the representative thumbnail for people.

Quicker Pages
A lot of data is downloaded when you click on a face from the homepage…it seems like a lot more than the number of thumbnails would entail. Then when you click on one of the thumbnails a lot more data is downloaded (this page also causes Firefox CPU usage to peg out at 100% as I mentioned in my previous post on Riya)….an option in preferences to set how many thumbnails are loaded on one page and how many photos are loaded on one page would be of great use to those of us who aren’t on the latest and greatest broadband connections.

Thumbnail “Unzoom”
When using the face recognition tools I find myself wanting to zoom out to the full photo on many of the thumbnails to help me identify who exactly that person is.

What Have I Uploaded?
A way to tell if I’ve already uploaded a photo or folder of photos to Riya. I originally uploaded a folder with about 1.2 gigs of folders to Riya, and when I went back to upload more, I couldn’t remember which folder I’d already uploaded. This would be easier if my photos were better organized in folders, but I use Picasa for all my organizational needs, making folder structure mostly irrelevant.

This Isn’t a Face!
I realize I noted this previously, but it’s aggravating me again. When using the bulk recognition tool, there needs to be an option for “This Isn’t a Face”. A good number of buildings and objects are recognized as faces by Riya. Of course the ideal way to do this would be to never mess up and call a building a face…

Larger Drop Downs
I have to do way too much scrolling in the drop down boxes to select names in the bulk recognition tool. Larger drop downs or a revamped interface would help address this. Additionally, if the face is near the bottom of the window the drop down isn’t fully contained in the viewable window space, so I have to scroll the whole page down and then scroll the drop-down box. The drop down should actually “drop” up in this circumstance.

It may not sound like it from all the “complaints”, but so far I find Riya to be very promising. It’s utility to me in it’s current state is limited because of it’s current inability to export to Picasa the metadata it generates, but I’m sure this feature will come along sometime. Riya’s core functionality is there, the featureset just needs to come around, and I have faith that will happen.

Tonight’s goal is to upload a bunch of new photos and see how much the 458 faces Riya has recognized in my current set helps in recognition of faces in the new set.

soalphaithurts, riya, photosearch

I’ve played with the alpha of Riya for a couple of hours now and here’s a few snippets of my experience.

First, the idea behind Riya is that it “tags” your photos with the names of the people in them. It does this automatically via facial recognition. It also uses text recognition to tag your photos with text on things that you took a photograph of (street signs, advertisements, etc). This whole idea rocks, so I’ve really been looking forward to the start of alpha testing.

It’s obvious that this is an alpha. But that’s ok. Here’s some critiques ripped straight out of my notes:

1. An estimated time left on the uploader applet would be sweet.
2. On the start page it tells you to click on “Unrecognized Faces” when in fact the link it’s referring to is entitled “Faces Unrecognized”.
3. Need a way to tell Riya that something it recognizes as a face, isn’t … or at least clarify the way this is done. Riya determined that some buildings and other objects were the faces of people. Humorously it also recognized a doll face in one of my photos.
4. Riya doesn’t like my face (very offensive!). I’m in many of the photos I uploaded but I had to go through dozens upon dozens of photos to get a good selection of samples whereas it popped up dozens upon dozens of photos of my cousin in the first few pages when training her face.
5. “Number of Training Examples for xxxx: xx” doesn’t seem to increment correctly when going through the training for an individual person (rather than through the bulk trainer). Several screens in a row the number will stay the same even if you select thumbnails as being representative of the individual, and then all of the sudden it will increment up to the correct number.
5. Way to correct name misspellings. My brain broke and I labeled a bunch of pics of my wife with her maiden name (we’ve been married for eight years, so I have no idea why I did this) so I went to Address Book to fix her name. Can’t do it, but can add a nickname, which didnt do anything. After doing this, I cannot display any photos for my wife. It shows that there are 27 photos with her in it, but when I click on her name or thumbnail it loads a lot of data into the web page and shows a lot of blank thumnail tiles, but after it’s done all those tiles disappear and it says cannot find any photos.
6. After clicking a thumbnail it loads a page full off medium sized versions of the photos in that album or that match that tag. This page makes firefox crawl and the CPU to peg at 100%.

Also a couple of general observations I noted down whilst playing around:

To upload photos to Riya you must download a java applet with which you select the photos you want to upload. I haven’t looked into it much, but this applet must do some pre-processing on the photos…maybe resizing them. It pegs my Athlon 64 3500+ at 100% usage for the whole time it’s uploading.

Which brings me to the next observation. In her invite email Tara Hunt stated that the uploader didn’t support Firefox yet. I had no problem using Firefox to upload my pics. I don’t see how you would have problem with any browser, since the uploader is a java applet you download and then run as a seperate client.

I also wondered if correcting mistakes in text recognition helped improve future accuracy?

riya, photosearch, soalphaithurts, riyarocks

I’m not much of an arguer in real life. I generally just accept what people say and if I disagree I just let them believe what they want to believe.

My internet persona is another beast altogether.

I can’t really pinpoint the reason behind that. All I know is if I see someone making a claim on the internet that doesn’t make sense or is just plain wrong, I cannot help but to commence correcting them.

The problem is that real-life arguments are much more fulfilling.

Here’s the phases of your typical ‘net argument.

Someone makes a false, illogical, or just plain stupid claim

Here’s a few examples of these type of claims:

“The ATI X1800XL is the best video card”

(Best in what way? Everything is best if you narrow the set of circumstances enough.)

“A Mac is the best choice for web development”

“Microsoft sucks.”

“Lost will be awesome for sweeps”

Naturally, being my compulsive, internet-arguing self, I reply in the negative to such claims. What comes next varies depends upon the savy of the person making the claim, whether or not they truly believe such a thing, or if they’re just a troll, but typically the next step is…

Defend false claim with anecdotal evidence or a point that sounds like it backs them up but really doesn’t

I hate to see this because some people truly don’t see the problem with it. If I can’t make them understand the battle is lost. (You never really win at internet arguing anyway)

“My brother just got an X1800XL and it’s SUPER FAST”

“My friend developed his l33t hacker website on his mom’s Mac and it was really easy.”

Notice the extensive use of friends and family to back them up. Like I care what they have to say.

“Windows crashed my PC and I lost all off the WoW hacks I’ve downloaded”

(One product on one PC in one particular set of circumstances does not a sucky company make.)

“Sweeps week helps them determine advertising revenues”

(Sure it does. Of course that doesn’t prove that Lost will rise from it’s recent suckiness for redemption during sweeps week. I’m sure the writers will try to put their most enthralling plot elements forward during sweeps. But then again, if they were able to produce awesome shows on demand why would they let the previous episodes of this season be so bad? While I realize “bad” is subjective, the audience as a whole has been disappointed.

This type of argument can be very frustrating. The person making the claim many times doesn’t see the subtle difference between their defense statement being correct and at the same time not proving their original point because the two points are so similar.)

I’ll try to point out to them the error in their ways by replying with a detailed response. This inevitably leads to…

Unfullfilling

This is where the discussion goes awry. There’s several ways your internet opponent can completly take all joy out of your life at this point.

“That’s just your opinion.”

Uh-huh

“Yeah, that’s what I’m saying!”

No you’re not. Stop trying to get on my boat of superiority.

“I like choclate chip cookies.”

Either realizing they have been beaten, or forgetting what they were talking about, many people choose to just pretend nobody ever said anything about anything.

“Let’s agree to disagree”

Let’s not, loser.

“See this benchmark shows that the X1800XL is fastest! I WIN! HAHAHAHAH!!! I got to go, later.”

Not only does the defense point not prove their point, they don’t stick around to hear what you have to say and go on forever believing they won the argument. Makes me so sad.

Mavromatic just got itself removed from my Bloglines subscription list.

“Why would you do such a thing to such a wonderful site, Dustin?”

They switched over to a partial-text feed with advertisements in EVERY POST.

My whole rationale behind using a feed reader is to streamline my information-gathering process. This sort of feed makes me go out to yet another site to get what the author wants me to read.

I gave Flock a try and it didn’t grab me right off. Since I’m pretty busy, I’m not going to take the time to really get into it.

I am now giving ecto a try. So far, it seems pretty nice.

In other news, I really need to devote more time to this blog. I’d like to get back up from the zero or so pageviews I get a day. To that end I’m going to implement some ideas stole straight from Steve Rubel. The use of ecto is job #1.

I’m giving Flock a try.

Kam VedBrat wants to know what we think of the UI in Windows Vista Beta 1.

I have mixed feelings about it. Most of what I think about it is summed up here. This is one of the larger internet forums around and is full of smart people.

I think you’d do yourself some good to read and even comment in that thread, Kam.

Gizmodo links to a working MiniSD WiFi card. This is amazing. In case you’ve never seen a MiniSD card before, one card is slightly bigger than your thumbnail.

My Audiovox SMT5600 has a MiniSD slot. I dont know how useful WiFi would be for it since there’s only one slot and taking out my 512MB card to put the WiFi card in would leave me with less than 20MB of on-board flash memory. Awesome, nonetheless.