Brlog - stories from a teen grad student

Another day in the life of blackrabbit. Okch.

I'm currently working on a phage for my research. Been trying to get the wild-type phage to work and finally today there's strong evidence that it's working! The phage ate all the E. coli on my plaque assay because I used a 1:1 ratio of phage to E. coli instead of a 10^-7:1 ratio :-/

Faure's 6th Nocturne. Haven't played it for ages. Haven't played the piano for ages either for that matter. At one point in time, I had at least 4/7ths of the song down being listenable¹, now I suspect I can only still play the first page -_-

I dunno.

In other news ... my wiki is pretty much done, only thing it needs is prettying up (something I'm not good at). I've already started using it to archive information XD

</br>

---

and better than ever? We shall find out.

I'm back in Taipei if anyone cares XD
I've nearly finished implementing my wiki/blog/site/thing. It's on it's second rewrite -_- ¹
I guess I think (and I hope) I improved my programming abilities over the summer hence the rewrite.

---

I rewrote a java remote desktop program to access the VNC settings for a mac remotely connected through SSH in about an hour and a half! hurrar! XD
Waste of time? XD

Purpose: Needed to access the school mac remotely using VNC.
Problem: VNC settings not turned on.
Available inventory: SSH is turned on.

Attempts at solving:
1. Google for ways to changing the mac remote settings through SSH.
2. Use a java remote desktop program I had written back in 2004 to change the VNC settings.
3. Write a new java remote desktop program to change the VNC settings.
So I couldn't find anything on the web about how to change the VNC settings through SSH (either that or my googleing skills suck now >.<). I had written a java remote desktop client&server back in 2004, but I couldn't find any copies. So I decided to write a new java remote desktop client&server. Even though I hadn't done socket programming for a while, I still remember the most essential classes needed. Took me about an hour to write, and 30 mins to debug XD
I guess I still haven't lost my knack for stuff I had previously learnt XD

Oh oh! and I only just found out that MouseListener now includes the mouseMoved(MouseEvent) and mouseDragged(MouseEvent) that used to be in MouseMotionListener (but MouseMotionListener isn't deprecated)?
Makes me wonder why they had it separate in the first place. Memory requirements?

Apparently someone unplugged my Ethernet cable XD

Btw, I'm in San Francisco at the moment for those that don't know XD
Still haven't got around to upgrading my blog ... I'm so busy here >.<

Had to write a program to take the intersection of two files, one with six million lines (~700MB), and the other two million lines (~250MB) of data. These were my attempts at writing the programs. I'm wondering if I'm getting weaker at programming :-/
(I could only go up to -Xmx1400m)

1st attempt: [Run: O(IO read time of 2 files),Mem: Asplode -_-]
A simple parsing everything into memory, extracting a primary key from each line, then using a HashSet to find the intersection.
Result: A very quick out of memory error.

2nd attempt: [Run: 2xO(IO read time of 2 files),Mem: Asploded too >.<]
Parse only the primary key of both files, put into a HashSet to find the intersection, then re-read the file a second time through to weed out the non-intersecting lines.
Result: A very slow out of memory error. Apparently even just two sets of primary keys was too much.

3rd attempt: [Run: O(IO read time of 3 files),Mem: ~1GB]
Parse the primary key of the shorter (first) file into a HashSet. During reading of the second file, start doing the intersection with each line so it won't take up extra memory. Put the intersecting items of the first Set into another Set so as to reduce the comparison time. Re-read the first file again, weeding out the ones that did not intersect.
Result: Enough memory left over to do some other calculations (an extra HashMap storing some number the size of the first file).

The first attempt was most obvious way, but the third was better than the second in both runtime and memory. I guess the second attempt as not as good an optimization as I had thought.

Man ... Bioinformatics requires dealing with tens of GB's of data, though usually you don't need to have everything in the memory at the same time.

[/rant]

So the moral of the story is ... ?

Being too tired from the flight, I somehow got infected by a spyware. Also being too sleepy, I wasn't in the right mind to get rid of it immediately, so the next time it showed up, I suspended it instead of killing it directly.

The fix was easy. Download new version of ad-aware, scan the folder where kbw.exe resides¹ and it'll remove it for you. I found out where it was executing from by using procexp.exe - a better task manager for windows than the original (it's one of the first things I put on any windows computer²).

DO NOT restart the computer right after it is removed.

What it turned out to be was an exe that hijacked the registry entry for exe files to execute their file. I didn't even know exe files had a registry entry telling it how to run. I had always thought they just ran. Anyway, I used Method 2 from this help page to fix it -- copied registry.exe to registry.com using windows explorer, then run that, then fix the default value of HKEY_CLASSES_ROOT\exefile\shell\open\command back to "%1" %*, and exe's started executing again.

Luckily I did not restart the computer immediately after I removed kbw.exe because I would not know what would've happened if windows could not load its startup executable files.

</br>

---