>The recent story of macro-evolution is available in current or recent evidence--the genomes of living creatures. Even if you completely ignore fossil evidence it is possible to find genetic commonalities between species that indicate common ancestors. With the fossil evidence it makes a very strong case.
As always the devil is in the detail. To get from one species to another (as currently determined by a pair of genetic pools) requires a viable path. Not just stepping stones but a real, viable path. For every protein, every gene, every process that is needed to support life.
The conventional dogma is that this occurs through a multiplicity of small changes (we have observed such small changes in real experiments - I have seen such data from my own lab) , each of which gradually moves the population (genetic drift). The problem is that whilst this is a cosy theory, there is little remenant of this. We should now have more species than at any time ever, and species should not be distinct but a continuum (if we take the slow steady mountain climb approach advocated by the likes of Dawkins). This is not something that makes sense but is very nice for putting very long timescales on things. Burst evolution, where dramatic environmental changes radically alter the genetic pressure on organisms whilst providing the circumstances to allow genetic change can be observed experimentally but, being unpredictable, throws all the happy 'slow steady climb' molecular clocks out of the window.
There are also distinct problems with establishing mechanisms for major events - evolution of sex, evolution of DNA duplication, evolution of completely novel genes etc. I'm not claiming these are insurmountable, but there are gaps in the set of mechanisms needed to demonstrate a full feasibility for an evolutionary origin.
And beyond establishing the possibility of a mechanism, to demonstrate that mechnism was THE cause is actually impossible. Evolution as a theory of origins is a faith, albeit one with a rational base that may or may not be orthogonal to any particular theological position. Evolution as a theory of relationships is powerful and useful in interpreting and understanding the way the world works now. To give a definition of micor and macro evolution.
Micro evolution: Directly obeservable changes to the composition (ie number of distinct variants) in a gene pool. Not to be confused with natural selection which does not increase the number of variants but just alters their relative proportion. We can do experiments on this.
Macro evolution: Hypothesis driven interpretation of existing data. This is not provable by experiment, but can be examined for consistency, though not correctness. A hypothesis which is not consistent is not correct. A hypothesis which is correct is consistent. A hypothesis which is consitent is not necessarily correct. This would include interpretataion of the fossil record (including genomic comparisons).
To say evolution is a done deal is far from the truth. Evolution is a useful tool for aiding understanding. For the agressively militant atheists who cling to it as a faith of origins, it is a crutch as much as a belief in creation is for a theist. From a scientific standpoint (and I am a scientist who makes use of evolutionary theory to aid understanding) I am quite content that I do not know how I got to be here. I am also aware that there are numerous consistent models for the formation of the earth which are incompatible, and numerous interpretations of the other data, again consistent but incompatible.
What should be interesting is to examine the result of the Tsunami in Asia in 2004 to see if fossil creation can be observed from such a cataclysmic action. I don't know if anyone is doing that...d
Likewise. I'll try X,Y,and Z three ways up if I know the shot is in there somewhere.
I went out to photograph a particular landscape. Took my compact digital (canon A60 - could do with an upgrade) and my MF TLR camera. With the digital I could play around and see exactly what was happening with the different filters/exposures. I then proceeded to shoot off a roll of film on the one scene with the MF camera. Seemed to work. I got a couple of usable frames from the MF that I wouldn't have got without the benefit of being able to preview the expected results with digital. Best of both worlds...d
The d200 allegedly compensates for that by capturing an identical 'blank' exposure to remoce sensor noise. It is capable of doing very long exposures (in excess of three hours) and will give analagous speed to film because of the lack of reciprocity failure. IME a standard 3MP consumer digital P&S gives analagous results to a film compact with 200ASA film in terms of quality. My Nikon F2 does better than the grain on Velvia with the prime lenses, but is somewhat soft with the Nikkor 80-200 f4.5 zoom. The modern range of dSLR (d2x etc.) require the very best in terms of glass otherwise you are not making the best of the available resolution. Likewise film - no point using Kodachrome 25 if you are using budget lenses. The pictures will still look crap. My Lubitel 166B (very cheap 6x6 camera) is pin sharp to digital print resolution (when examined with a loupe) for infinity shots at 12" square. I'll get some 20" square prints back soon and see what they are like. Film is not dead, as long as it is MF or better. I would imagine that once I get the high performance dSLR that the MF camera will not get used much...d (saving for a d200, only GBP1100 to go;-( plus a decent lens for it..) 3 Nikon SLR, bunch of lenses, two compacts digital, two compact 35mm. Still can't get landscapes horizontal.
It was pretty smudgy here as well. I dragged the camera and tripod out though and grabbed a few shots. Hopefully they will be a bit clearer than my binocs. (100 asa film, 30 secs F4 200mm lens should do it..)..d
Aw c'mon, thats just the summary score card. A real cricket score card is the ball by ball account of the match from which everything can be reconstructed (short of the actual field placements.)
Just scoring a cricket game takes a lot more skill than just counting runs...d
I think you missd the point. Waterhouse's religion plays a key role in how he relates to the others in the society around him. In a time when there is turmoil and religious conflict, being part of an outspoken minority has its own perils and direclty shapes the behaviour of W. in his pursuits and choices.
There are many complex social threads running through the book. It was a lot deeper than I was expecting and very entertaining for someone with a sense of history and place. (He is talking my language as a chemist, and the developement of London during the scientific enlightenment of the seventeenth and eighteenth centuries is fascinating as it's echos are still apparent in much of the modern city. Some parts are contrived but are forgiven as they are necessary to bridge the gaps in the narrative.
At first I didn't realise that it was a trilogy so got to the end of the first volume and felt somewhat cheated as there were more loose ends than I could count still unaccounted for.
For those who are still looking for a plot, you are wasting your time. There are minor plots and subplots, but it is a historical narrative, along the lines of War and Peace or similar. There is no major 'Grand Design' a la Cryptonomicon where the story rotates around teh ultimate treasure hunt. In the Baroque Trilogy the story centres araound the characters and their relationships with the times and implicitly how those times relate to us in terms of the knowledge and developments that we take for granted.
Quicksilver was slow. It is not a pulp fiction page turner. It is not a thriller. But taken on it's terms as a narrative of characters and times rather than of a plot, it is an excellent work. The pace is jsut right and I am looking forward to the next two volumes (just ordered The Confusion).
What was least believable? Jack Shaftoe's Eastern European trek and much of the travel. The timescale is far too compressed. Apart from that, it doesn't rankle too much, and is far better than most of what passes for literature these days...d
We had ca 25% of our deathstars fail inside two years. Out of about 30 drives we have had to change 7 or 8 due to failure. That is a failure rate we don't see with other brands.
Come on, surely you have some percieved NEED before you spalsh out on a PC for home. Get software to meet that need, your need. Not software to create needs you never knew you had.
sheesh.. some wierd people out there. I can understand people asking if there is good software to do X but 'I've got a PC, what shoudl I do with it?' takes the biscuit, even for slashdot...d
I was learning Norwegian and found that it was easier to be understood if I mumbled than if I spoke clearly. We listen and read in context so skip the odd word here or there. Why do you think it is so hard to proof read typos out of a piece of text in your native tongue?..d
The object with bioinformatics is to get answers, not necessarily to create tools (unlike many more traditional applications). Some history from someone on the peripherary.
In teh beginning was data and a bunch of programmers in different places with different skills asking different questions.
Toolkits were developed with different aims in mind:
BioPerl is a programmers toolkit in perl with strong links to the ensembl project (heck, it IS the Ensembl project).
EMBOSS (not represented at teh biohackathon unfortunately) is an application oriented toolkit in C.
Likewise biojava, biopython etc.
Each of these projects has its own strengths and weaknesses but has reached an appropriate level of maturity where it has become apparent that working together is essential for the best future development of all teh projects.
The aim of the biohackathon is to bring together sufficient coders for sufficient time to allow a lot of the minor (and most are minor) differences and interoperatability issues with each particular toolset to be ironed out so that they can each leverage each others strengths.
With the dispersed nature of the coders (many, many continents) it is fantastic that O'Reilly and Electric Genetics have sponsored this get together. Those of us who have to answer new questions on a day to day basis are extremely gratefull for the toolkits provided, and the ability to hack them to suit the task in hand.
There is a lot of iterative engineering going on. If we were to sit down and desigh to the nth degree then we would still be waiting for a standard model of DNA in a computer..
Instead we write poor code that gets some results. Others pick it up and make it better because they need to. I can think of thousands of lines of code I didn't need to write because of these projects. This equals results faster equals new therapies faster.
I too am a unix oriented person.
OS X is very pretty but has a few annoyances..
1. Apps seem to want to hang temporarily for no apparent reason (spinning disc for many seconds)
2. I have my home area on a Linux machine, mounted via NFS and authenticated through NIS:
- Terminal doesn't log me in. I have to log in as a local user then SU to my network ID. This sucks rocks.
- X is not quite happy on starting from teh application icon for a networked user.
Other than that it is great. I like fink/apt and it seems to work just fine. It provides a solid name/target for ISV's to target whilst allowing all my open source stuff to run just fine. A definite plus.
>>At present I get my colleagues to prototype their databases with something like filemaker and then reimplement in postgresql.
>Why? Surely one of the point-and-click interfaces to PostgreSQL would make more sense?
I prefer emacs/vi as my interface to db design (and a whiteboard and coloured pens) but they are using (typically) MacOS 9 and are happy using the interfaces filemaker provides. It also means they screw up their own system rather than mine.
In an earlier life I was using Access as a reporting engine for a postgres DB because its rapid report generation capabilities are easy to use and far ahead of the point and click interfaces.
Using the appropriate tool isn't nececssarily just finding the tool that could do the best job. It is a case of tool, operator and situation. In this case using Fielmaker on MacOS is a good tool for them and it readily ports (data and structure) to a more powerful DB system.
I am currently building a molecular biology database for sequence analysis. By biology standards it is quite small (dealing with about 100K entries for sequences, a handful of tables and it isn't terribly complex.
However, I am adding results from similarity searches (it takes time to do these) and piling them into the database. I estimate approximately 3 million records within the next month.
So far it is up to about 400K records and seems to be behaving very nicely.
This is postgres 7.0.3 (stock version with SuSE 7.1).
I intend to migrate to 7.1.3 or later in due course because of the record size limitation in pre 7.1 versions. It remains to be seen how well it will cope with text searches of 3 million records (about 1.5secs with 250K entries in the table) or whether I have to do some keyword/prefix tree munging to speed things up.
If the whole thing falls over then I'll have to look at why and whether I need bigger hardware (currently on 2xPIII/833 1Gb RAM) or a bigger database management program but that would mean money...
At present I get my colleagues to prototype their databases with something like filemaker and then reimplement in postgresql.
Show me a corporation that gives it secretaries a
new PC, a windows install CD, the CD's for Word etc...
They don't, IT do it all for them and give them a preconfigured out of the box solution. The same happens with Linux. I have taken a group of students (life sci with no 'geek cred'), sat them in front of a preconfigured box and let them get on with it. Star Office, no problem. Other apps, no problem.
Preconfigured, ready to roll with the apps one needs.. That is what appears on your desktop, not a bunch of CD's and so on.
Apart from that, the author does appear to be a couple of years behind the times. RH 7.0 was a breeze to install. Insert CD (no other boot disk needed), follow the instructions and bingo, a workstation ready to roll. Of course you need to know some technical details like your user name and the network settings, but when did MS ever set those up (cue gMindRead 0.1)..
I personally know someone who has seen this happen for real. They were stationed on the Falklands at the time (Early 90's).
Fly a helicopter slowly towards the beach, just high enough for the little darlings to not scatter and they will indeed watch it until they topple over backwards.
Just because one PR officer says it aint so, doesn't mean that those who have actually seen it are lying. Apparently it is quite hysterical (and it is sheep, not penguins that can't get up when they fall on their backs).
OK, I have downloaded the genome, indexed it and have it available for my users.
The latest full release of EMBL (63) weighed in at about 4.7 Gb compressed. This took me about 30 hours to download.
GPL'd tools are available. Checkout EMBOSS for a start, BioPerl, BioJava, bioPython, and BioXML, all linking in with a common biocorba interfaces, and many more besides.
I run my bioinformatics service with a minimum of commercial software (only one commercial package which I am soon replacing with EMBOSS, and several non-open packages. The majority are open to some degree.
Needless to say it is based on Unix systems (IRIX/Linux in my case).
This is an interesting point of view though it contains some factual errors, eg. take point 33. It lists eight ways a file can be identified with the system. He claims only three are allowed and five are forbidden for no reason. This is wrong. Five are allowed and three are forbidden. The five allowed are
One name
many names
in one directory
In many directories
sharing a directory with other files
The disallowed are
No name
sharing the same name with other files
Being in no directory
I suppose one could also add 'existing but nonexistent' to the list of disallowed ones as well.
An information-centric view of data is very interesting and something I have been thinking about for my own field (bioinformatics). Having a 'browser' that reconfigures depending on the datatype being viewed and its relationships with other data types is a fascinating idea. I can think of some ways to get close to a 'fileless' method of interaction that I would like to pursue but do not have the time.
Given that a microarray (glass slide method) is basically a glass microscope slide that has been treated by a few chemicals (takes a few mins with a few old glass jars as supports) and the real expense comes in spotting the array (needing a machine to accurately spot 40K spots on a normal microscope slide) it should be possible to do small arrays (a few hundred spots) at home. Then you need a reader which can detect the spots. Using old tech you could use radiation and xray film, then scan it on your normal PC scanner.
Bingo.. instant home microarrays (but 200K for kit would get it done a lot quicker and more accurately.. wonder if you could build an array spotter from lego mindstorms..)
So, say you could change something cosmetic about yourself genetically for a reasonable price. For example, what if a virus were available that triggered a whole-body genetic mutation, and the end result was a change in your genetic hair color?
So this raises an interesting question that I've wanted an answer to for some time. [side note: I am not a biologist, nor do I play one on the Net, so please excuse me if this is a dumb question].
One of the much touted advantages of genetic engineering is the ability to cure genetic problems in living humans. This is distinct from altering the genetic code in cells that will go on to form a viable human foetus.
So, say I have some genetic disease caused by an unfortunate sequence in my DNA. Assume we know what replacement sequence would cure this problem. On an engineering level, how would I go about making the change in every cell in my body? This is what I would have to do, right? Is this an area where nanotechnology and genetic engineering meet? Or could genetically-modified viri really perform this task?
I assume that now we are closing in on getting detailed genetic information about humans, people are starting to think about how gene therapy might be applied in practice. Does anyone have anything they can share with us on this subject?
It is not a dumb question by any sense of the word.
A genetic disease can have two major effects.
Over production of a protein that is harmful
Under production of a protein (or production of a flawed protein) that is needed (eg haemophilia, which is a defect in blood factor VIII or factor IX)
The first is very hard to treat as you do have to hit every cell in the body.
The second is much easier adn fortunately much more the common scenario. You do not need to correct every cell, just persuade enough cells to produce the right protein (usually by putting in some extra DNA that has the correct coding sequence) that the balance is restored.
Interestingly the same technique can be used to immunise people. Instead of injecting protein, inject DNA that then produces teh protein where it can raise a bigger immune response.
Does it work?
Unfortunately the technology is nottrivial and the understanding of how this works is far from complete which is a scientists way of saying yes and no.
There has been limited success in some trials though humans don't seem to respond as well as the animal models.
I should mention that I am somewhat of an outsider in this field, I am not as up to date as once I was so a more up to date biologist can probably give you a better answer.
Slashdot Mirror
>The recent story of macro-evolution is available in current or recent evidence--the genomes of living creatures. Even if you completely ignore fossil evidence it is possible to find genetic commonalities between species that indicate common ancestors. With the fossil evidence it makes a very strong case.
..d
As always the devil is in the detail. To get from one species to another (as currently determined by a pair of genetic pools) requires a viable path. Not just stepping stones but a real, viable path. For every protein, every gene, every process that is needed to support life.
The conventional dogma is that this occurs through a multiplicity of small changes (we have observed such small changes in real experiments - I have seen such data from my own lab) , each of which gradually moves the population (genetic drift). The problem is that whilst this is a cosy theory, there is little remenant of this. We should now have more species than at any time ever, and species should not be distinct but a continuum (if we take the slow steady mountain climb approach advocated by the likes of Dawkins). This is not something that makes sense but is very nice for putting very long timescales on things. Burst evolution, where dramatic environmental changes radically alter the genetic pressure on organisms whilst providing the circumstances to allow genetic change can be observed experimentally but, being unpredictable, throws all the happy 'slow steady climb' molecular clocks out of the window.
There are also distinct problems with establishing mechanisms for major events - evolution of sex, evolution of DNA duplication, evolution of completely novel genes etc. I'm not claiming these are insurmountable, but there are gaps in the set of mechanisms needed to demonstrate a full feasibility for an evolutionary origin.
And beyond establishing the possibility of a mechanism, to demonstrate that mechnism was THE cause is actually impossible. Evolution as a theory of origins is a faith, albeit one with a rational base that may or may not be orthogonal to any particular theological position. Evolution as a theory of relationships is powerful and useful in interpreting and understanding the way the world works now. To give a definition of micor and macro evolution.
Micro evolution: Directly obeservable changes to the composition (ie number of distinct variants) in a gene pool. Not to be confused with natural selection which does not increase the number of variants but just alters their relative proportion. We can do experiments on this.
Macro evolution: Hypothesis driven interpretation of existing data. This is not provable by experiment, but can be examined for consistency, though not correctness. A hypothesis which is not consistent is not correct. A hypothesis which is correct is consistent. A hypothesis which is consitent is not necessarily correct. This would include interpretataion of the fossil record (including genomic comparisons).
To say evolution is a done deal is far from the truth. Evolution is a useful tool for aiding understanding. For the agressively militant atheists who cling to it as a faith of origins, it is a crutch as much as a belief in creation is for a theist. From a scientific standpoint (and I am a scientist who makes use of evolutionary theory to aid understanding) I am quite content that I do not know how I got to be here. I am also aware that there are numerous consistent models for the formation of the earth which are incompatible, and numerous interpretations of the other data, again consistent but incompatible.
What should be interesting is to examine the result of the Tsunami in Asia in 2004 to see if fossil creation can be observed from such a cataclysmic action. I don't know if anyone is doing that.
Likewise. I'll try X,Y,and Z three ways up if I know the shot is in there somewhere.
..d
I went out to photograph a particular landscape. Took my compact digital (canon A60 - could do with an upgrade) and my MF TLR camera. With the digital I could play around and see exactly what was happening with the different filters/exposures. I then proceeded to shoot off a roll of film on the one scene with the MF camera. Seemed to work. I got a couple of usable frames from the MF that I wouldn't have got without the benefit of being able to preview the expected results with digital. Best of both worlds.
>>Oh, and my 35mm SLR doesn't need batteries. I can't say the same for any digital imaging device...
..d
>you nincompoop - yes it does. all SLR's, film or digital, need batteries, to cock/release the mirror and shutter.
That's why I can take pictures quite happily after removing the batteries on my F2 or FE2? Why many old cameras don't have any kind of battery at all?
Sunny 16 rule anyone?
The d200 allegedly compensates for that by capturing an identical 'blank' exposure to remoce sensor noise. It is capable of doing very long exposures (in excess of three hours) and will give analagous speed to film because of the lack of reciprocity failure. IME a standard 3MP consumer digital P&S gives analagous results to a film compact with 200ASA film in terms of quality. My Nikon F2 does better than the grain on Velvia with the prime lenses, but is somewhat soft with the Nikkor 80-200 f4.5 zoom. The modern range of dSLR (d2x etc.) require the very best in terms of glass otherwise you are not making the best of the available resolution. Likewise film - no point using Kodachrome 25 if you are using budget lenses. The pictures will still look crap. ..d ;-( plus a decent lens for it..)
My Lubitel 166B (very cheap 6x6 camera) is pin sharp to digital print resolution (when examined with a loupe) for infinity shots at 12" square. I'll get some 20" square prints back soon and see what they are like. Film is not dead, as long as it is MF or better. I would imagine that once I get the high performance dSLR that the MF camera will not get used much.
(saving for a d200, only GBP1100 to go
3 Nikon SLR, bunch of lenses, two compacts digital, two compact 35mm. Still can't get landscapes horizontal.
"We the people, for the people" is a socialist manifesto.
Discuss
It was pretty smudgy here as well. I dragged the camera and tripod out though and grabbed a few shots. Hopefully they will be a bit clearer than my binocs. (100 asa film, 30 secs F4 200mm lens should do it..) ..d
The frequency histogram is pretty flat, suggesting a rotational cipher. Might try taking each line as an increased increment in the rotation.. ..d
Aw c'mon, thats just the summary score card. A real cricket score card is the ball by ball account of the match from which everything can be reconstructed (short of the actual field placements.)
..d
Just scoring a cricket game takes a lot more skill than just counting runs.
I think you missd the point. Waterhouse's religion plays a key role in how he relates to the others in the society around him. In a time when there is turmoil and religious conflict, being part of an outspoken minority has its own perils and direclty shapes the behaviour of W. in his pursuits and choices.
..d
There are many complex social threads running through the book. It was a lot deeper than I was expecting and very entertaining for someone with a sense of history and place. (He is talking my language as a chemist, and the developement of London during the scientific enlightenment of the seventeenth and eighteenth centuries is fascinating as it's echos are still apparent in much of the modern city. Some parts are contrived but are forgiven as they are necessary to bridge the gaps in the narrative.
At first I didn't realise that it was a trilogy so got to the end of the first volume and felt somewhat cheated as there were more loose ends than I could count still unaccounted for.
For those who are still looking for a plot, you are wasting your time. There are minor plots and subplots, but it is a historical narrative, along the lines of War and Peace or similar. There is no major 'Grand Design' a la Cryptonomicon where the story rotates around teh ultimate treasure hunt. In the Baroque Trilogy the story centres araound the characters and their relationships with the times and implicitly how those times relate to us in terms of the knowledge and developments that we take for granted.
Quicksilver was slow. It is not a pulp fiction page turner. It is not a thriller. But taken on it's terms as a narrative of characters and times rather than of a plot, it is an excellent work. The pace is jsut right and I am looking forward to the next two volumes (just ordered The Confusion).
What was least believable? Jack Shaftoe's Eastern European trek and much of the travel. The timescale is far too compressed. Apart from that, it doesn't rankle too much, and is far better than most of what passes for literature these days.
We had ca 25% of our deathstars fail inside two years. Out of about 30 drives we have had to change 7 or 8 due to failure. That is a failure rate we don't see with other brands.
But I don't know why!!
..d
Come on, surely you have some percieved NEED before you spalsh out on a PC for home. Get software to meet that need, your need. Not software to create needs you never knew you had.
sheesh.. some wierd people out there. I can understand people asking if there is good software to do X but 'I've got a PC, what shoudl I do with it?' takes the biscuit, even for slashdot.
There are hydro plants that do this. They use off peak electric to pump water up hill then use that water to generate electricity at peak times.
..d
There is one in Snowdonia that does this.
Works fine for all languages..
..d
I was learning Norwegian and found that it was easier to be understood if I mumbled than if I spoke clearly. We listen and read in context so skip the odd word here or there. Why do you think it is so hard to proof read typos out of a piece of text in your native tongue?
And to throw a spanner in the neatness theory, one isn't optically active at all (Glycine) and one is not strictly an amino acid (its an imino acid).
And for those who like R and S notation, one is neither, and two are S whereas all the others are R. (Sulphur is a higher precedence than Oxygen).
Wrong answer..
The object with bioinformatics is to get answers, not necessarily to create tools (unlike many more traditional applications).
Some history from someone on the peripherary.
In teh beginning was data and a bunch of programmers in different places with different skills asking different questions.
Toolkits were developed with different aims in mind:
BioPerl is a programmers toolkit in perl with strong links to the ensembl project (heck, it IS the Ensembl project).
EMBOSS (not represented at teh biohackathon unfortunately) is an application oriented toolkit in C.
Likewise biojava, biopython etc.
Each of these projects has its own strengths and weaknesses but has reached an appropriate level of maturity where it has become apparent that working together is essential for the best future development of all teh projects.
The aim of the biohackathon is to bring together sufficient coders for sufficient time to allow a lot of the minor (and most are minor) differences and interoperatability issues with each particular toolset to be ironed out so that they can each leverage each others strengths.
With the dispersed nature of the coders (many, many continents) it is fantastic that O'Reilly and Electric Genetics have sponsored this get together. Those of us who have to answer new questions on a day to day basis are extremely gratefull for the toolkits provided, and the ability to hack them to suit the task in hand.
There is a lot of iterative engineering going on. If we were to sit down and desigh to the nth degree then we would still be waiting for a standard model of DNA in a computer..
Instead we write poor code that gets some results. Others pick it up and make it better because they need to. I can think of thousands of lines of code I didn't need to write because of these projects. This equals results faster equals new therapies faster.
We are not writing wordprocessors here.
..d
I too am a unix oriented person.
OS X is very pretty but has a few annoyances..
1. Apps seem to want to hang temporarily for no apparent reason (spinning disc for many seconds)
2. I have my home area on a Linux machine, mounted via NFS and authenticated through NIS:
- Terminal doesn't log me in. I have to log in as a local user then SU to my network ID. This sucks rocks.
- X is not quite happy on starting from teh application icon for a networked user.
Other than that it is great. I like fink/apt and it seems to work just fine. It provides a solid name/target for ISV's to target whilst allowing all my open source stuff to run just fine. A definite plus.
..d
>>At present I get my colleagues to prototype their databases with something like filemaker and then reimplement in postgresql.
>Why? Surely one of the point-and-click interfaces to PostgreSQL would make more sense?
I prefer emacs/vi as my interface to db design (and a whiteboard and coloured pens) but they are using (typically) MacOS 9 and are happy using the interfaces filemaker provides. It also means they screw up their own system rather than mine.
In an earlier life I was using Access as a reporting engine for a postgres DB because its rapid report generation capabilities are easy to use and far ahead of the point and click interfaces.
Using the appropriate tool isn't nececssarily just finding the tool that could do the best job. It is a case of tool, operator and situation. In this case using Fielmaker on MacOS is a good tool for them and it readily ports (data and structure) to a more powerful DB system.
..d
I am currently building a molecular biology database for sequence analysis. By biology standards it is quite small (dealing with about 100K entries for sequences, a handful of tables and it isn't terribly complex.
However, I am adding results from similarity searches (it takes time to do these) and piling them into the database. I estimate approximately 3 million records within the next month.
So far it is up to about 400K records and seems to be behaving very nicely.
This is postgres 7.0.3 (stock version with SuSE 7.1).
I intend to migrate to 7.1.3 or later in due course because of the record size limitation in pre 7.1 versions. It remains to be seen how well it will cope with text searches of 3 million records (about 1.5secs with 250K entries in the table) or whether I have to do some keyword/prefix tree munging to speed things up.
If the whole thing falls over then I'll have to look at why and whether I need bigger hardware (currently on 2xPIII/833 1Gb RAM) or a bigger database management program but that would mean money...
At present I get my colleagues to prototype their databases with something like filemaker and then reimplement in postgresql.
..d
Show me a corporation that gives it secretaries a new PC, a windows install CD, the CD's for Word etc...
They don't, IT do it all for them and give them a preconfigured out of the box solution. The same happens with Linux. I have taken a group of students (life sci with no 'geek cred'), sat them in front of a preconfigured box and let them get on with it. Star Office, no problem. Other apps, no problem.
Preconfigured, ready to roll with the apps one needs.. That is what appears on your desktop, not a bunch of CD's and so on.
Apart from that, the author does appear to be a couple of years behind the times. RH 7.0 was a breeze to install. Insert CD (no other boot disk needed), follow the instructions and bingo, a workstation ready to roll. Of course you need to know some technical details like your user name and the network settings, but when did MS ever set those up (cue gMindRead 0.1)..
Fly a helicopter slowly towards the beach, just high enough for the little darlings to not scatter and they will indeed watch it until they topple over backwards.
Just because one PR officer says it aint so, doesn't mean that those who have actually seen it are lying. Apparently it is quite hysterical (and it is sheep, not penguins that can't get up when they fall on their backs).
The latest full release of EMBL (63) weighed in at about 4.7 Gb compressed. This took me about 30 hours to download.
GPL'd tools are available. Checkout EMBOSS for a start, BioPerl, BioJava, bioPython, and BioXML, all linking in with a common biocorba interfaces, and many more besides.
I run my bioinformatics service with a minimum of commercial software (only one commercial package which I am soon replacing with EMBOSS, and several non-open packages. The majority are open to some degree.
Needless to say it is based on Unix systems (IRIX/Linux in my case).
- One name
- many names
- in one directory
- In many directories
- sharing a directory with other files
The disallowed areI suppose one could also add 'existing but nonexistent' to the list of disallowed ones as well.
An information-centric view of data is very interesting and something I have been thinking about for my own field (bioinformatics). Having a 'browser' that reconfigures depending on the datatype being viewed and its relationships with other data types is a fascinating idea. I can think of some ways to get close to a 'fileless' method of interaction that I would like to pursue but do not have the time.
Oh for a few students..
Well--
Given that a microarray (glass slide method) is basically a glass microscope slide that has been treated by a few chemicals (takes a few mins with a few old glass jars as supports) and the real expense comes in spotting the array (needing a machine to accurately spot 40K spots on a normal microscope slide) it should be possible to do small arrays (a few hundred spots) at home. Then you need a reader which can detect the spots. Using old tech you could use radiation and xray film, then scan it on your normal PC scanner.
Bingo.. instant home microarrays (but 200K for kit would get it done a lot quicker and more accurately.. wonder if you could build an array spotter from lego mindstorms..)
..d
So this raises an interesting question that I've wanted an answer to for some time. [side note: I am not a biologist, nor do I play one on the Net, so please excuse me if this is a dumb question].
One of the much touted advantages of genetic engineering is the ability to cure genetic problems in living humans. This is distinct from altering the genetic code in cells that will go on to form a viable human foetus.
So, say I have some genetic disease caused by an unfortunate sequence in my DNA. Assume we know what replacement sequence would cure this problem. On an engineering level, how would I go about making the change in every cell in my body? This is what I would have to do, right? Is this an area where nanotechnology and genetic engineering meet? Or could genetically-modified viri really perform this task?
I assume that now we are closing in on getting detailed genetic information about humans, people are starting to think about how gene therapy might be applied in practice. Does anyone have anything they can share with us on this subject?
It is not a dumb question by any sense of the word.
A genetic disease can have two major effects.
The second is much easier adn fortunately much more the common scenario. You do not need to correct every cell, just persuade enough cells to produce the right protein (usually by putting in some extra DNA that has the correct coding sequence) that the balance is restored.
Interestingly the same technique can be used to immunise people. Instead of injecting protein, inject DNA that then produces teh protein where it can raise a bigger immune response.
Does it work?
Unfortunately the technology is nottrivial and the understanding of how this works is far from complete which is a scientists way of saying yes and no.
There has been limited success in some trials though humans don't seem to respond as well as the animal models.
I should mention that I am somewhat of an outsider in this field, I am not as up to date as once I was so a more up to date biologist can probably give you a better answer.