Two problems with that 1) The p2p network they have developed / using be will restricting access to the UK by using Geo-IP (http://sourceforge.net/projects/geoip/).
2) The only Doctor Who episode likely to be broadcast during the trial period is the Christmas Special.
Your biggest problem (or if you are lucky, your sucessors) is maintaining the hardware to be able to read the WORM media.
I work in Document Management and have seen problems in the past with Optical drives, mainly with support in newer OS's and hardware; trying to find SCSI cards that work in new machines with old (8 bit) connectors, drivers for modern OS's and also replacement drives.
I have had great fun in the past when a customers almost new (9 months old) drive failed only to be told that they had stoped making them and could not replace or repair the drive and there was no upgrade path as they had decided after 6 months that the design was a dead end.
Their only suggestion was to look in the small adds for someone selling an 'old' drive.
One other to make sure is that the format you are storing them in is an open standard (e.g. TIFF). I could tell you stories of that would give you nightmares about vendor 'lock-in'.
We had to cut huge blocks of stone from a quary in Wales, drag it half across the country to Salisbury Plain, put them up (by hand!). Then we had to wait till summer solstice to see if it ran correctly.
You don't even need to recalibrate, just modify the search engine to use fuzzy techniques.
See www.historicaldirectories.org for an example.
Historical Directories is a digital library of local and trade directories for England and Wales, from 1750 to 1919. It contains high quality reproductions of comparatively rare books, essential tools for research into local and genealogical history.
1. Open regedit 2. Create the registry key HKEY_CURRENT_USER\Software\Google\Google Desktop 3. Create a string value entry named "install_dir", the value of which should be the desired install path 4. Create a string value entry named "data_dir", the value of which should be the desired data path (where GDS will keep the indexing data). 5. Run the GDS installer.
Hypothetically someone could modify the code to Mozilla / Firefox or any other open source program and put up the executable for download:)
No I don't secure my personal data by merely watching network traffic, I do run new apps in a 'sandbox' of a system under VMWare with external connections disabled and examine what is happening with the software (file calls, registry, TCP / UDP traffic). So far I have not used GDS to index my IE Cache I don't see the need because a) I don't use IE for the majority of browsing 2) I don't see the need when I can just go to Google.com although that will mean that Google.com then has an idea of my surfing habits or at least my searching habits.
The index is already compressed. They would only be able to monitor external, public, websites via a sending URL method which would be of limited effectiveness.
Your can install it on any drive you like, however, it store the index in your profile directory (normally c:\Documents and Settings\YOURUSERNAME\Local Settings\Application Data\Google\Google Desktop Search\) and wants to make sure that you have more than enough space to store the index.
I am sure that somewhere there must be a hack to move your documents and settings folder to another drive.
Full text indexing algorithms are public knowledge. I haven't seen anything that jumps out as some previously unknown technique. It takes my documents and builds a set of inverted file tables that I can then search. Lots of products take this approach, most of the time the only differences are in the number of supported file types and the search techniques (boolean, proximity, range, phrase, wildcard, fuzzy, quorum) available to the user.
They store the index of your data in your profile folder under documents and settings. The 1GB value is large enough to store an index approximately 10GB of data (need to run some more tests to find out proper index to corpus ratio).
It was probably hard coded because a) it is beta software and b) not asking questions that may confuse the average user is seen as user friendly. The average user only has one drive, C: because that is how it was shipped from PC World.
Well, as the index is about 10% of the size of the corpus (total data indexed) I would expect to notice a large chunk of outgoing traffic from my network.
Although PDF support is not explicitly listed I am getting PDF documents in the results when I search. It seems that there is some PDF support in there, it is just not advertised.
It is a full text search engine. It indexes the content of the documents. It is not going to be every word, the majority of full text search engines have a list of stop / noise words which they do not index becuase they are so common.
A work collegues car has an LPG car, fitted from new by the manufacture. Turns out he will still have to pay because the LPG system was not fitted by an (the only) approved supplier
Slashdot Mirror
Two problems with that
1) The p2p network they have developed / using be will restricting access to the UK by using Geo-IP (http://sourceforge.net/projects/geoip/).
2) The only Doctor Who episode likely to be broadcast during the trial period is the Christmas Special.
Well she was, in a way. In H2G2 Arthur had been trying to chat Trillian up before Phil (AKA Zaphod) whisked her away from under his nose.
Your biggest problem (or if you are lucky, your sucessors) is maintaining the hardware to be able to read the WORM media.
I work in Document Management and have seen problems in the past with Optical drives, mainly with support in newer OS's and hardware; trying to find SCSI cards that work in new machines with old (8 bit) connectors, drivers for modern OS's and also replacement drives.
I have had great fun in the past when a customers almost new (9 months old) drive failed only to be told that they had stoped making them and could not replace or repair the drive and there was no upgrade path as they had decided after 6 months that the design was a dead end.
Their only suggestion was to look in the small adds for someone selling an 'old' drive.
One other to make sure is that the format you are storing them in is an open standard (e.g. TIFF). I could tell you stories of that would give you nightmares about vendor 'lock-in'.
Hit highlighting on the image is nothing new products like ZyIMAGE (http://www.zylab.com/) have been doing exactly this for 8-10 years now.
They have already announced MSLinux (see http://www.mslinux.org/), it was supposed to have shipped Nov 2003, looks like it has been delayed again :)
Punches, you had it easy!
We had to cut huge blocks of stone from a quary in Wales, drag it half across the country to Salisbury Plain, put them up (by hand!). Then we had to wait till summer solstice to see if it ran correctly.
You don't even need to recalibrate, just modify the search engine to use fuzzy techniques.
See www.historicaldirectories.org for an example.
Historical Directories is a digital library of local and trade directories for England and Wales, from 1750 to 1919. It contains high quality reproductions of comparatively rare books, essential tools for research into local and genealogical history.
From the GDS newgroup
1. Open regedit
2. Create the registry key HKEY_CURRENT_USER\Software\Google\Google Desktop
3. Create a string value entry named "install_dir", the value of which should be the desired install path
4. Create a string value entry named "data_dir", the value of which should be the desired data path (where GDS will keep the indexing data).
5. Run the GDS installer.
Hypothetically someone could modify the code to Mozilla / Firefox or any other open source program and put up the executable for download :)
No I don't secure my personal data by merely watching network traffic, I do run new apps in a 'sandbox' of a system under VMWare with external connections disabled and examine what is happening with the software (file calls, registry, TCP / UDP traffic). So far I have not used GDS to index my IE Cache I don't see the need because a) I don't use IE for the majority of browsing 2) I don't see the need when I can just go to Google.com although that will mean that Google.com then has an idea of my surfing habits or at least my searching habits.
The index is already compressed. They would only be able to monitor external, public, websites via a sending URL method which would be of limited effectiveness.
Your can install it on any drive you like, however, it store the index in your profile directory (normally c:\Documents and Settings\YOURUSERNAME\Local Settings\Application Data\Google\Google Desktop Search\) and wants to make sure that you have more than enough space to store the index.
I am sure that somewhere there must be a hack to move your documents and settings folder to another drive.
Yes the searching takes place on your computer. It builds an index of your data in your profile folder under c:\documents and settings.
It appears to be using pretty standard relevence ranking, rather than any special google ranking techniques.
Full text indexing algorithms are public knowledge. I haven't seen anything that jumps out as some previously unknown technique. It takes my documents and builds a set of inverted file tables that I can then search. Lots of products take this approach, most of the time the only differences are in the number of supported file types and the search techniques (boolean, proximity, range, phrase, wildcard, fuzzy, quorum) available to the user.
They store the index of your data in your profile folder under documents and settings. The 1GB value is large enough to store an index approximately 10GB of data (need to run some more tests to find out proper index to corpus ratio).
It was probably hard coded because a) it is beta software and b) not asking questions that may confuse the average user is seen as user friendly. The average user only has one drive, C: because that is how it was shipped from PC World.
Well, as the index is about 10% of the size of the corpus (total data indexed) I would expect to notice a large chunk of outgoing traffic from my network.
Although PDF's are not listed as supported, it is returning some PDF documents when I search.
Although PDF support is not explicitly listed I am getting PDF documents in the results when I search. It seems that there is some PDF support in there, it is just not advertised.
It is a full text search engine. It indexes the content of the documents. It is not going to be every word, the majority of full text search engines have a list of stop / noise words which they do not index becuase they are so common.
The problem with Quantum voting is that you will change the result just my measuring it.
A work collegues car has an LPG car, fitted from new by the manufacture. Turns out he will still have to pay because the LPG system was not fitted by an (the only) approved supplier