Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Silkroad -- Ignore unavailable to you. Want to Upgrade?

To: Kachina who wrote (462)	8/6/1999 6:25:00 PM
From: Frank A. Coluccio	Respond to of 626

Wavelets... from my limited understanding of them, were what I was referring to. Glad you raised the distinction. Maybe someone else who has more background on this topic than either of us can contribute something further. I've also used the term wavelets in recent posts to connote something else, probably in error. I used the term to describe very narrow line width derivatives of commercially available lambdas (which are usually spaced at 50 GHz or 100 GHz. Such wavelengths would be harnessed through a second stage of beam division, to much finer granularities. These "wavelets," I've been know to state, would be lower order wavelengths with much narrower spacing requirements and far lower information carrying capabilities, which could be used for lower speed applications. Like, say, a mere 1 to 5 Gb/s or so for building risers (backbones), or even less where desktop applications are being supported, instead of the usual 10 or 20 Gb/s currently being assigned to carrier grade DWDM lambdas.

To: Kachina who wrote (462)	8/9/1999 7:19:00 AM
From: EPS	Read Replies (1) \| Respond to of 626

Hi Kachina, try to do this: make a function equal to 1 on [0,1], zero outside, second step divide [o,1] in two pieces and define a second function to be 1 in [0,1/2], -1 in [1/2,1], and keep repeating the subdivisions. you get an orthonormal system of functions (the Haar system).if you try to write functions in terms of series of these Haar functions you need to know the coefficients (just like in Fourier series you need to know the Fourier coefficients or Fourier transform). How many coefficients do you need to get a good approximation of your function/or better how many coefficients do you need so that the error in your approximation is small enough for your needs? compression then means algorithms to select those coefficients efficiently. given the simple nature of the process it is easy to compress and decompress wavelets..note that from one original function (the mother) you are able to generate all the other Haar functions. There are of course other more sophisticated wavelet systems. fractals are objects that have fractional dimension*. sometimes they can be generated by selfsimilarity process associated with a construction similar to the one discussed above. the idea is that each level of resolution you can only see structurally the same that you saw at previous levels. in this fashion complicated worlds can be created with a very simple mechanism. the hope is that sometimes chaotic systems (ie those nonlinear systems where a small perturbation of initial conditions can produce great perturbations) can be described in these terms. of course one can only hope so much but it does happen sometimes.. hopes this helps.. Victor