Monday, November 15, 2010

Omar Edgardo Sanchez Noya, rest in peace, 2nd part.


  1. It turns out that I knew him, recognized is face right away, a big picture of him was posted outside the library. It was a long time ago, but I used to eat in the cafeteria, and in order to skip the long queue I'd trick anybody into believing some excuse for me skipping the queue such as being in a hurry or something like that. That's how I met Omar, he was a short person, slim and had a characteristic hairstyle: a really messy one. He let me skip the queue several times, he used to carry two bags, one for his books and one for his clothes as he used to do physical training. I don't remember what we talked about while waiting our turn, but he was a nice guy. I remember inviting him his meal in exchange for helping me save at least an hour in the queue. He was a loner too as far as I'm concerned, he had friends but most of the time I saw him alone. Many people congregated today including myself outside the library to show our condolences to his family. I don't know why am I so sad, I barely knew him, I think the sole thought of my own inevitable death is what really brings me down. I wish I wasn't this sentimental, it's very burdensome.

  2. Super-scalar architectures are those that implement two or more execution units. The key for this technique is that some instruction require more processing than others so, the microprocessor may implement a mechanism for characterizing an instruction or group of instructions, and reorder them in a way that most of the parallel execution units are busy most of the time. This kind of architecture is said to apply dynamic reallocation, when the compiler does this reallocation we're talking about static reallocation. Some post optimization enhancements can reallocate data-independent instructions in the gaps required to synchronize dependent instructions.

  3. Hierarchical representations of data can be very effective retrieving data but the main shortcoming lays in the fact that relations many-to-many are hard to handle, this is when relational relations come into play. I remember Miguel saying something about OLAP, but I didn't quite get it.

  4. The FAT file system is implemented in devices such as flash memories due to its simplicity, everything is stored in a big table and write/read accesses are minimized even if data gets highly fragmented...mmm that's a bit contradictory isn't it?

  5. The Achilles heel of unix systems is that they can't cope very well when resources are very scarce, we were told, but I don't buy that either, I need to see it with my own eyes, I can tell about my linux desktop distribution but it should be a very different picture when regarding servers.

  6. We revisited java interfaces, talked about RMI, and how it's implemented. I really like the teacher Nava giving us programming lessons, he declares himself as not being software-savvy (his passion is hardware) but If you ask me, he is better than most of the teachers that give us programming classes.

  7. Data-warehouses are a very neat concept, I wonder how are they implemented. This is the kind of stuff that attracts my attention the most, as Oscar Herrera says...torture the data 'till it speaks.

No comments:

Post a Comment