Algorithms

Re: Lexical Analysis Useful for Other Things as Well?

4-Aug-09 6:13

Richard Andrew x64 wrote:
by tokens that a parser can understand

IMO a computer does not understand tokens, all it does is execute the rules you make up for it to identify tokens.

And yes, lexers and parsers can be used by more than compilers; their goal may be to calculate something (interpreters), to draw something (graphing calculators), to proof something (math theory building), etc.

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

The quality and detail of your question reflects on the effectiveness of the help you are likely to get.
Show formatted code inside PRE tags, and give clear symptoms when describing a problem.

modified on Tuesday, August 4, 2009 12:24 PM

N a v a n e e t h4-Aug-09 7:57

N a v a n e e t h

4-Aug-09 7:57

Completely agree with Luc.

Richard Andrew x64 wrote:
is it possible to use parsing technology for broader things than just compilers?

Aren't parser technologies used for other jobs than compiler designs?

Richard Andrew x64 wrote:
I've just begun studying parsing and lexical analysis, and I find it fascinating.

It is all fascinating stuff. But it takes a long time and effort to come up with a decent working parser. I am working on an open source parser to parse Malayalam[^] language text and do grammar checking and it is taking long time than I expected.

Smile | :)

Navaneeth

How to use google | Ask smart questions

Re: Lexical Analysis Useful for Other Things as Well?

harold aptroot6-Aug-09 2:58

harold aptroot

6-Aug-09 2:58

I would have replied earlier but I didn't have a good example, but now I came across one. Scientific chemical names. They need to be parsed (since their structure is non-trivial), but that can't be called compiling Smile | :)

Re: Lexical Analysis Useful for Other Things as Well?

N a v a n e e t h6-Aug-09 18:10

N a v a n e e t h

6-Aug-09 18:10

Have a look at http://www.wolframalpha.com/[^]. They parse the input nicely and using for various purposes.

Navaneeth

How to use google | Ask smart questions

Re: Lexical Analysis Useful for Other Things as Well?

Eddy Vluggen8-Aug-09 7:13

Eddy Vluggen

8-Aug-09 7:13

Richard Andrew x64 wrote:
I wonder if parsing algorithms could be used to make a computer understand, say, the rules of a game, or the design of a spacecraft.

Yes and no. There is always ambiguity in language. One needs to be able to reason before you can "understand" a text. We (as humans) deduce meanings, interpret a text, or even skip words. It get's worse if the text contains things like sarcasm, pictures or figure of speech.

Richard Andrew x64 wrote:
In other words, if we can represent concepts by tokens that a parser can understand, is it possible to use parsing technology for broader things than just compilers?

Sounds like Protege[^] Smile | :)

Parsing a text is one thing. Reasoning about it's content is a step up. Understanding a (short) text[^] - is even more difficult.

The computer is mightier than the pen, the sword, and usually the programmer.

Backus-Naur Form Terminology

Richard Andrew x643-Aug-09 13:36

Richard Andrew x64

3-Aug-09 13:36

I have trouble with the concept of a "production" in BNF.

Take, for instance, the following BNF snippet:

D  := '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' | '8' | '9'

Proper BNF terminology calls the things on the RIGHT side of the :=
"productions."

However, since the thing on the left of the := is made up of the things on the right, doesn't it make more sense to call the D the production?

Confused | :confused:

Luc Pattyn3-Aug-09 15:23

3-Aug-09 15:23

would it help to call "D" the product, i.e. the result of the production?

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

The quality and detail of your question reflects on the effectiveness of the help you are likely to get.
Show formatted code inside PRE tags, and give clear symptoms when describing a problem.

Richard Andrew x643-Aug-09 15:28

Richard Andrew x64

3-Aug-09 15:28

That might help, but that nomenclature does not appear in the literature.

It just seems like such a backwards use of the word. And when words are used in alternative ways like that, it interferes with my understanding of the material.

Luc Pattyn3-Aug-09 15:36

3-Aug-09 15:36

Hi,

let me try again.
IIRC the righthand side is called a "production rule", not a "production".
Does that make more sense?

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

The quality and detail of your question reflects on the effectiveness of the help you are likely to get.
Show formatted code inside PRE tags, and give clear symptoms when describing a problem.

Any algorithm for allocation

Richard Andrew x643-Aug-09 15:39

Richard Andrew x64

3-Aug-09 15:39

Oh, well, yes, that does make more sense.

I guess the paper I have been reading used a lazy abbreviation of the proper term.

That sheds a whole new light on things. Thanks. Smile | :)

Sabari MD31-Jul-09 1:03

Sabari MD

31-Jul-09 1:03

hai

i need to fill the blank space with certain images of diffarent sizes.images should be well placed in the space next to previous one.

sabarikuttan

Sabari MD
Application Developer
Veloxit Info Solutions

harold aptroot31-Jul-09 1:28

harold aptroot

31-Jul-09 1:28

But what is "well placed"? Do you mean you want to do Optimal Rectangle Packing with irregular sizes and an additional constraint on the shape of the result? (you will not like the solution..)

Sabari MD31-Jul-09 1:45

Sabari MD

31-Jul-09 1:45

Yes my rquirement is

I have rectangular free space .

i need to fill it with images of diffarent sizes.

Sabari MD
Application Developer
Veloxit Info Solutions

Luc Pattyn31-Jul-09 1:39

31-Jul-09 1:39

That is known as the knapsack problem. Look it up!

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

The quality and detail of your question reflects on the effectiveness of the help you are likely to get.
Show formatted code inside PRE tags, and give clear symptoms when describing a problem.

Fingerprint Image Quality Algorithm

Moreno Airoldi1-Aug-09 1:13

Moreno Airoldi

1-Aug-09 1:13

As Harold said, it really depends on what you mean by "well placed".

If you just treat all images as rectangles, then a greedy algorithm (see knapsack problem, as Luc suggested) surely is the simplest and fastest solution.

If you go for that, remember you should manage situations where there is no solution (for example if you want ALL images placed in your blank space: it may not be possible). You may also have to go through different placing patterns in order to find a solution, so a one-go greedy algorithm may not be sufficient.

2+2=5 for very large amounts of 2
(always loved that one hehe!)

Member 90279629-Jul-09 21:23

Member 902796

29-Jul-09 21:23

Hi
i want 2 implement a project in matlab tittled as 'Fingerprint Image Quality Classification'

where we get an image and classified it as wet/oily , dry and normal image .
Can anybody like 2 share his work or help me out?

Regards

Re: Fingerprint Image Quality Algorithm

Fatbuddha 13-Aug-09 1:17

Fatbuddha 1

3-Aug-09 1:17

Can you define for me wet/oily , dry and normal image?

You have the thought that modern physics just relay on assumptions, that somehow depends on a smile of a cat, which isn’t there.( Albert Einstein)

Re: Fingerprint Image Quality Algorithm

Member 9027963-Aug-09 18:20

Member 902796

3-Aug-09 18:20

wet and oily image is one in which ridges are overlapped by dark circle , give impression as black ink has been spotted on fingerprint .

Dry image is one in whcih some pixels from ridges are rubbed

normal is good quality image

Re: Fingerprint Image Quality Algorithm

Fatbuddha 13-Aug-09 23:49

Fatbuddha 1

3-Aug-09 23:49

So its been a while since I did image processing and it was not in that field.
But my first thought would be if I can great a sort of pattern for each state and then try to correlate this.
Or you could try to eliminate the finger in the image and check the rest.
Something like that.
I don't know if this is helpful, because as I said its been a while.

So good luck!

Cheers

You have the thought that modern physics just relay on assumptions, that somehow depends on a smile of a cat, which isn’t there.( Albert Einstein)

Barrel distortion correction

GloriaWiktoria27-Jul-09 10:51

GloriaWiktoria

27-Jul-09 10:51

Hi,

I know that I should not ask a general question but I spend whole week looking for any solution and I am just fed up. I need some easy guide (I am not mathematician) for the barrel distortion correction algorithm.

Thanks in advance.

Re: Barrel distortion correction

Luc Pattyn27-Jul-09 11:36

Tiny Encryption Algorithm Help?????

27-Jul-09 11:36

I suggest you read some of these hits[^].

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

The quality and detail of your question reflects on the effectiveness of the help you are likely to get.
Show formatted code inside PRE tags, and give clear symptoms when describing a problem.

ChiSmile26-Jul-09 9:50

ChiSmile

26-Jul-09 9:50

Hello all,
Am trying implement TEA in my application. So far I have been able to encrypt a table with eleven colums. But i can only decrypt three of the colums. I dont seem to understand what the problem is because even the colums have similar datatype i.e. nvarchar.
why would it chose to decrypt only three colums. Below is the code where i passed the arguement.

For Each rw As DataRow In dt.Rows

                   decdata = New TEA()
                   deRecDataFnam = decp.Decrypt(rw.Item("First Name"), ned)
                   deRecDataLnam = decDatalab.Decrypt(rw.Item("Last Name"), ned)
                   deRecDataSec = decdata.Decrypt(rw.Item("Social Security Number"), ned) --> can decrypt
                   deRecDataPh = decdata.Decrypt(rw.Item("Phone Number"), ned)
                   deRecDataCon = decdata.Decrypt(rw.Item("Contact Address"), ned)
                   deRecDataDOB = decdata.Decrypt(rw.Item("Date of Birth"), ned) --> can
                   deRecDataMar = decdata.Decrypt(rw.Item("Marital Status"), ned)
                   deRecDataBlud = decdata.Decrypt(rw("Blood Group"), ned)
                   deRecDataGeno = decdata.Decrypt(rw.Item("Genotype"), ned) --> can
                   deRecDataGen = decdata.Decrypt(rw("Gender"), ned)

               Next

the encryption and decryption code can be found here
http://allmysocial.net/post/Tiny-Encryption-Algorithm-(TEA)-in-Visual-BasicNET.aspx[^]

I can paste the algorithm if u want me to.
Thanks a lot.

Re: Tiny Encryption Algorithm Help?????

Luc Pattyn27-Jul-09 11:38