The Ultimate English Proof

Some ... well, most ... scientific journals have English proofing as an official stage in pre-publication life of your paper. The need for this stage is obvious -- it is becoming increasingly rarer when authors are native in English. In fact, this guy is not a native English spe... hm, writer.

Got this in email today:

The yellow marker is the revision made by the English proofer. In fact, this is a PDF comment. No actual comment, just the marker. Ve-e-e-e-ry unclear on what I should do about it. So, I had to send a reply email asking for clarification.

In the meantime the verdict is: the ultimate English proofer is a guy who tells you WHAT to write in your papers rather than a guy who fixes spelling and grammatical errors. The idea that I could alternate several writing styles and allow for some frivolity here and there ... is now shattered.

Take care ResearchGate. Hello Google Scholar!

Not particularly interested in social networking when in comes to research. Do not even remember how my ResearchGate account was created in the first place.

For awhile I have have this TODO item that says: Fix my ResearchGate publication list. Got on the website today and tried to do it properly. Particularly, the site only had less than 30 of my publications which was definitely not right and had to be fixed.

ResearchGate has the BitTex import option. Tried it today. DOES NOT WORK. Yes, it will import author names and titles, ... and years but not the important stuff like conference names, pages, etc. I also hear that even after you import something ResearchGate might change the names later on. Not something I would like to see on my list.

So. Cancelled my ResearchGate account. Today. Goodbye ResearchGate!

Now, there is Mendeley, CiteULike, and all these other options. I guess CiteULike might be about the same as Google Scholar but just in case I got on Google Scholar first. And, to my own surprise, it came up with the correct number of my publications -- 132 sounds about right.

Granted the site is ugly, this automatic feature alone has made my mind. Next I have to figure out how to connect my source code at GitHub and slides at SlideShare -- naturally you cannot upload the actual PDFs due to copyright, right? -- to the articles on my Google Scholar. But at least I do not have to import anything.

Hello Google Scholar!

What does IEICE I-SCover Cover?

Really busy right now. Had to find my previous publication to put into References and came across this:

For those who do not know, this is the I-SCover -- the new search service commissioned by IEICE last year. Could not let this pass. In the "Releated Metadata" Section, not a single keyword is actually related to the paper. Note that the actual keywords are right there. However, the presence of Related Metadata section hints at backend mapping between my keywords and what they consider "related". Which means that my paper is listed in some completely unrelated section of their metadata.

You can see the page itself at this link.

This particular idea is definitely worth shredding.

On Cheap Action Video Cameras

Finally got around and purchased an action video camera. Had to return the first one and buy another one. These are the lessons I learned in the process.

The first one I got was Looxcie LX2. Its amazon page was relatively short yet I did see that it could capture at most 480p videos and went for it anyway, thinking that ... well, that it might still work out. Smaller videos might actually be a better thing..... or so I thought. Besides, it boases of a nice smartphone application that you could use with the camera.

Wrong I was. Not only video quality was bad pixelwise, the encoding ... I am guessing implemented in software, was very poor. During abrupt scene changes ... which by definition are frequent in wearable settings ... the frame would go berserk for about a second during which boxes in various places on the screen would display junk. Completely unsuable.

Also, the feature that made my mind -- the 30s clip feature, where it would dump into the file the LAST 30 seconds of the buffer when you press a button --- would only work when configured from smartphone, which takes Bluetooth pairing, connection inside the application ... all amounting to 30-60s preparation before you can use the feature. Note that after that you have to walk around with your smartphone.

Returned the thing. Got Contour ROAM2 for only 30% on top of the price of the first one (I know, I know, the price is so low only becuase the company went bust recently ... not gloating or anything...). Did not get it yet, but I already know all the features, which I can now list in full, having learned my lesson:


(1) anything below 1920x1050 is junk. Other names for it are 1080p or Full HD. You will still get blurry transisions on abrupt changes, but very brief and not very obvious.

(2) several kinds of mounts would be nice

(3) rotating lens! is a major feature. This way you can adjust your horizon properly, even if you mounted it awry.

(4) some way to tell where you are aiming. Contour shoots a lazer line for aiming.

(5) resilience to elements. Contour is waterproof. Does not mean that you can swim with it but at least I can carry it in the rain without warrying too much.


Yep. I have my first wearable camera now.

Stupid Glass and the Real Google Policy

Heard about the "do not be evil" policy in Google? Given the Google Glass, the actual policy seems to be the one from the Phineas and Ferb -- "That's a google, googles don't do much, you know..."

Meet Google Class:

Basically, you have a small monitor stuck just above your eye line. And you get a camera and a button near your temple. That's all. Are you happy with the device? Lots of people seem to be happy. Lots of people hurry to state that this is a new mode of computing altogether. Is it really?


Meet Augmented Reality:

AR is simple in principle. Video is local, metadata is provided by a remote service over the network and is mapped onto the video as a faint overlay.

Google Class could be a new mode of computing if it made the whole eye socket of the glasses, or possibly both, into a monitor and project overlay over it. Only then the Glass would actually be fun to have and use. Otherwise, like I said, it is just a tiny screen glued to the tip of your baseball cap.

True Social Collaboration

Now sure how you the reader define online social collaboration, but this is pretty much how it is done today. Using the old western cliche ... there is only two kinds a' people -- those who are in line for collaboration and those who are let into the ant farm and are already collaboratin'.

Twitter #hashtag Algorithm Has to Be Reverse-Engineered .. Now

I think I finally got Twitter #hashtags. Based on the visual below which I took from here and digested into this brief analysis.

I think there are 2 problems, both internal to Twitter:

(1) their marketing people are bad at visualizations -- for example, the same bottleneck in the chart above is used twice thus complicating understanding -- stopping short of stating a deliberate intent to confuse the reader; I know that visualizations are hard, but this is Twitter and this is a very simple topic -- how hard can it be?

(2) Twitter marketing is completely decoupled from its technical department. It looks like the marketing is a bunch of slogan writers (read: dreamers) while the technical department works on their algorithms in complete isolation. Many important conditions for your tweet to show up in a hashtag stream have been discussed before -- none of them are shown in the chart.


Also about the chart, what does "ARE YOU GOING TO USE PROMOTED PRODUCTS TO SURFACE YOUR MESSAGE ON TWITTER" mean? Am I stupid? Have I missed some meeting on social media terminology? Note that this is not the only example of bogus terminology on the chart -- just look at it closely.


You know what? After this chart I got a BAD ITCH to reverse engineer the #hashtag algorithm. This has been done before -- the crawling, I mean. Since Twitter is wide open I can just traverse a bunch of hashtags ... and separately people and see who shows up and who does not.

Maybe then I will get an idea why most of the tweets that show up in #googlemapsapi #gis are either stupid exclamations or completely irrelevant (adding value?) posts. For example, today in #gis I saw a post of some student who got an E (not A!) in geography ... I guess we're talking middle or high school.

Back to work.