being civil never hurts and it makes your arguments much more effective when presented to a wider audiance.
By being downright childish and petty all you do is weaken the argument that you are putting forward as your attitude becomes associated with your reasoning.
Actually I have thought about the whole subject of how language impacts upon coding (every now and then I pick up a basic language text from something from a totally different linguistic family - usually chinease or Japanease but I'm meaning to get around to russian and maybe look at some american indian languages).
And while I agree with other posters who have said that a language does not completely shape the way people view the world I do think that the subtlties associated with any given linguistic family do have a strong impact upon the eligance of the solution.
This has two clear implications for language design:
1) In 1984 newspeak is a deliberatly restricted set of language components designed to make it impossible for people to express certain opinions. This can be taken forward into language design to prevent people from constructing "unsafe" constructs - i.e. removing GOTO's as a command, possibly hiding dynamic memory allocation (ala Java) or maybe enforcing extreamly strong static typing. These changes often make coding a solution a bit more difficult but can lead to more provable and reliable code.
2) You can take the other approach and allow many different constructs all with similar but subtly different meanings. This is predominalty the approach which Perl takes with its mixture of very high level constructs (grep, et al.) and very low level (interupts, if's,etc...). This leads to a language which is very easy to express yourself in but has the problem of being quite hard to read.
None of this of course has touched upon any of the functional languages (haskal, gopher, PROLOG and to a certain extent SQL) which try to do away with any notion of considering the flow of control and move simply to specifiying the required solution.
In short I think that the languages which the developer of a program has an appreciation of does shape the solution which they eventually arrive at - and the language which they are working in determines how clean the implemenation of that solution is.
does anyone actually know what the license costs for Oracle on Linux actually is? If this is a datawarehouse application then I'm assuming that you can't use one of the low end single user licenses.
Whenever I have looked at Oracle the software costs have dwarfed the hardware costs.
Well I work for an e-commerce consultancy in the UK and we have developed a payment gateway which is available either as a boxed product of as a managed service. Its currently in use in serveral contries as the back end card processor for several major Telco e-commerce offerings.
Integration is via a very simple thin API (using Baltimore libraries to protect card details in transmission if its a remote site) - we have language bindings in C, Perl, Java, TCL (I think), ColdFusion and a COM object for ASP's on win32 platforms.
Assuming that you want a managed service then we charge is a flat 1% of transaction revenue (on top of whatever your aquirer charges) and provide secure online authorisation, settlement, credit and reversal functionality - the ability to process a large number of different card types (MC, Visa, Amex, JCB, Delta, etc...).
In addition the bureau provides fraud detection functionality (card velocity checking, blacklists, max/min order values etc...)
If you're interested then take a look at
http://www.commercelink.co.uk/ (appologies for the ugly site - its being rebuilt)
I think that the (consumer based) sites which are going to make money online are those which fall into one of the following categories:
1) they provide a superior way of performing some sort of task (i.e. online banking or news feeds)
2) they provide an equivalent service but their market segment is composed of "tech savvy" people - e.g. computer components suppliers.
3) their site provides an integrated discussion/community area and goods purchasing forum (i.e. online building suppliers with an associated DIY site).
4) they can genuinely offer savings over traditional stores, something which (once you factor tax breaks) only a few online businesses actually achieve.
Some sites fall into several of these categories - for example I would say that Amazon fulfilled almost all of them:
* you can search for specific titles rather than having to physically hunt through the shelves * the ratings and suggestions systems make it easy to locate new books * I suspect a significant amount of their business is in computer literature.
I'm a very long way from convinced that an online shop which is just "online" will ever be particularly useful for a small business. Consumer hubs and price comparison services make it very difficult for a small business to actually generate very much interest in their site let alone generate many sales, and what many/.'s will forget is that web sites are only cheap if you know how to build one yourself.
Like I said you have to add some value with your site if you want it to make money - and I think that this is what the VC community is just waking up to - of course I'm not a VC so maybe that's just an impression from the media.
Other people in this discussion have talked about tipping sites or selling very low ticket goods, the problem is that presently (in a distributed environment) its not very easy to process those small amounts. Online payments really need to be authorised against something or the fraud risks become enormous - now this is either by using a traditional bank account which incurs minimum authorisation charges (typically around 5p which makes truly micro payments impossible) or against some form of online stored value system (either developed as part of your site or externally i.e. Beanze).
However micropayment solutions are still a long way from pervasive and I'm not sure if consumers are really interested - people like to have a single account that they keep track of and having to occasionally charge a separate online account seems like something of an overhead - perhaps if the account was set up with a direct debit so that your online "charge" was transparently topped up when required it would be easy to use - but then you run the risk of huge bills being generated if the architecture is compromised.
In short I'm not convinced that micropayment based revenue is going to be practical within the next few years - I'd be happy to hear different but I just don't see the social or economic hurdles evaporating.
depends on what you mean - OpenMarket have a goods type called micropayments in their Transact product. Basically a lump payment made to a server which then handles pay per view URL's on behalf of the merchant (well near enough for the sake of this conversation). I'm 100% sure they patented that system.
However the term micro-payment is usually applied to any transaction below £1 (or $) so I don't think you could patent the ability to process a small amount of money (unless you are a major bank of course *grin*)
actually one of the major problems with micro payments over the net is the overhead of processing these - you either have a choice of not auth'ing the card (in which case you could see fraud rates as high as 70%) or you have a processing cost which could be greater than the ammount in question.
Personally I want to get one as soon as they come out - I've been wondering about buyind a DVD player for a while and I'm just gonna hold out for the PS2 - that way I get the games console part of it for practially free!
I don't belive the problem here is anything to do with moral "should it" "shouldn't it" be taxed discussions - likewise I don't belive that this is legislation which applies solely to web based trade. Its much more to do with the practicality of deciding where the tax liability is incurred and then applying the tax in that location - mostly this problem is down to deciding where the transaction takes place.
e.g. Company based in USA, server hosted in Bermuda, customer in Canada.
Whose tax rules do you use? Currently people seem to be tending towards using the location of the server for any service (i.e. online content) and for the location of the customer for physical purchases. (Well this is the situation in europe anyway). But this is far from clear, it wouldn't suprise me if at some point people start to incure taxes at all sites involved!
Under current systems companies account for their VAT liability - but with a serious increase in cross boarder trade those companies need to be VAT registered in multiple locations if they ship above a certain revenue to those locations.
The only real alternative is to make every individual responsible for handerling thier own VAT (sales tax) liability and applying the tax laws local to the customer - I don't belive that this is going to be practical anytime soon due to shear volume!
Hence, if anybody really wants to tax trade over the net there needs to be a clear descision about where transactions take place - and nobody wants to do that because all pure net businesses will instantly "relocate" for the tax advantages!
Actually I can personally contradict this - I wrote my final year project (in 1998) up on a Mac SE. Now while this is an old machine - it is still perfectly caperble of word processing meaning that I only had to hunt for lab space in order to print.
On the otherhand I have also used both Staroffice and Wordperfect 8 under linux (neither extensively). These also seem to work fine for writing documents for printing elsewhere (never did get my printer properly configured).
But in addition I could have written code and surfed the net (had I not been studiously writting up my project that is)!.
In essence for the average user one WP is much like another, basic style support and some form of graphical import is all thats really essential to the base user.
Interesting viewpoint - however if I had that number of machines at home I'd probably want a rack mount too!
Just cause somebody wants their kit neat doesn't make them anal... imho properly organised kit is usually easier to access than a rats nest of wires.
As for advocating messy code in in OS enviroments I think you are wrong - clean code is easy to maintain - the advantage of OS is to have multiple people working on a project - if the code is "intricate" then there are fewer people who can join in.
Make stuff simple - make it clean. But of course that is just personal preference.
From what I understand a breakup is not punishment - its a way of removing the monopoly. From my laymans point of view I would expect punishment to be in the form of either a HUGE fine or of some forced concessions to competitors.
Its one of the things I liked about my university course... over my four years I covered Ada, PROLOG, Gopher, Z80, 68000, C/C++ and a host of other languages.
But more importantly we covered data structures, alorithm analysis and design looked at software lifecycles and examined lots of specialist areas (real time systems, os design, compiler design, AI's/computer vision, cryptography, etc...)
Nowadays I'm working at an e-commerce consultancy... most of my time is spent integrating with payment gateways some of which are written in Cobol on a VAX and some of which are have CORBA idl mappings.
I'm usually working in languages/enviroments which I haven't specifically seen before the start of the week - but so far I have always seen the underlying concept before.
I guess thats the difference a good 4 year course makes.
Yup all fair points... I wasn't suggesting this as a certaintly just one possability. More to the point I wasn't suggesting that there was a disproportionate amount of geniuses in those populations... merely that they where not being fully tapped as part of the labour force.
Moreover my main point was simply to agree with Rob that there are likely to be a number of unpredictable shocks to the IT/computing industry.
Given that 3rd world tech shops have been a concern for a while I am perfectly happy to accept that they are unlikely to be that shock... but there are plenty of unknowns left out there.
I really don't belive that post is worth a 5 - but whats worrying is that lots of this readership seems to. This is the same sort of self agrandesment as we get with any Linux story - look at this objectivally guys if you don't you run very serious financial risks.
What Rob actually said wasn't that there was going to be a huge increase in the supply of programmers but that there was almost certain to be a structural shock to the industry... he specifically said that he could not predict the source of this shock.
This could be that:
a) More problems are solved by pre-packed solutions - hence less need for custom solutions or from a sys-admin point of view maybe vast leaps forward are going to be made in reliability.
b) A new language comes out which lower quality programmers can use to achive equivalent results.
c) Large amounts of new labour become available - look at all the companies which have experimented with outsourcing their projects to 3rd world techies... these guys are just as bright, work just as hard (or harder) and cost fractions of a western worker.
The likelyhood remains that if this industry (IT et al) where to remain structurally the same then salaries for techies will level off and then fall in the medium term due to increased supplies of new graduates (of which I am one).
But look at the longer term and the chances of some random inovation reducing the numbers of us which are required or much more likely changing the TYPE of techie that is required is pretty good.
We are still an infantile industry - demand is high prices are volotile but can it really last indefinatly.
I doubt it and I'm certainly not going to bet my future/pension and morgage on it.
just thoughts
Tom
Re:nice, slashdot effect
on
GNUstep 0.6.0
·
· Score: 1
A couple of people have suggested mirroring the site... but that sounds like quite a lot of work if you want to expand it accross all stories posted.
howabout a/. proxy (storyproxy.slashdot.org)? Would it be possible to set it up so that only URL's (and thier sub directories?) actually posted within main stories would be cached?
It would have a lower maintenance than trying to mirror every site that get posted on/. but I guess it would still cost quite a few $$$ in bandwidth and machines.
Besides think what it would do to the polls if 30-40% of users all went throught he/. proxy?
I think the comment may have been more directed at people who include all of their major code in GUI event handlers rather than having the handler call another procedure.
If all you do is propogate calls to another procedure (probably in a separate object file) then your maintenance becomes much simpler.
Slashdot Mirror
Well actually I did do that with my car - but on the other hand it is a 1976 MG Midget *grin*
fair point
being civil never hurts and it makes your arguments much more effective when presented to a wider audiance.
By being downright childish and petty all you do is weaken the argument that you are putting forward as your attitude becomes associated with your reasoning.
Actually I have thought about the whole subject of how language impacts upon coding (every now and then I pick up a basic language text from something from a totally different linguistic family - usually chinease or Japanease but I'm meaning to get around to russian and maybe look at some american indian languages).
And while I agree with other posters who have said that a language does not completely shape the way people view the world I do think that the subtlties associated with any given linguistic family do have a strong impact upon the eligance of the solution.
This has two clear implications for language design:
1) In 1984 newspeak is a deliberatly restricted set of language components designed to make it impossible for people to express certain opinions. This can be taken forward into language design to prevent people from constructing "unsafe" constructs - i.e. removing GOTO's as a command, possibly hiding dynamic memory allocation (ala Java) or maybe enforcing extreamly strong static typing. These changes often make coding a solution a bit more difficult but can lead to more provable and reliable code.
2) You can take the other approach and allow many different constructs all with similar but subtly different meanings. This is predominalty the approach which Perl takes with its mixture of very high level constructs (grep, et al.) and very low level (interupts, if's,etc...). This leads to a language which is very easy to express yourself in but has the problem of being quite hard to read.
None of this of course has touched upon any of the functional languages (haskal, gopher, PROLOG and to a certain extent SQL) which try to do away with any notion of considering the flow of control and move simply to specifiying the required solution.
In short I think that the languages which the developer of a program has an appreciation of does shape the solution which they eventually arrive at - and the language which they are working in determines how clean the implemenation of that solution is.
just random musings.
Tom
does anyone actually know what the license costs for Oracle on Linux actually is? If this is a datawarehouse application then I'm assuming that you can't use one of the low end single user licenses.
Whenever I have looked at Oracle the software costs have dwarfed the hardware costs.
Well I work for an e-commerce consultancy in the UK and we have developed a payment gateway which is available either as a boxed product of as a managed service. Its currently in use in serveral contries as the back end card processor for several major Telco e-commerce offerings.
Integration is via a very simple thin API (using Baltimore libraries to protect card details in transmission if its a remote site) - we have language bindings in C, Perl, Java, TCL (I think), ColdFusion and a COM object for ASP's on win32 platforms.
Assuming that you want a managed service then we charge is a flat 1% of transaction revenue (on top of whatever your aquirer charges) and provide secure online authorisation, settlement, credit and reversal functionality - the ability to process a large number of different card types (MC, Visa, Amex, JCB, Delta, etc...).
In addition the bureau provides fraud detection functionality (card velocity checking, blacklists, max/min order values etc...)
If you're interested then take a look at
http://www.commercelink.co.uk/
(appologies for the ugly site - its being rebuilt)
or drop me an email
cheers
Tom
Nothing new or novel below
I think that the (consumer based) sites which are going to make money online are those which fall into one of the following categories:
1) they provide a superior way of performing some sort of task (i.e. online banking or news feeds)
2) they provide an equivalent service but their market segment is composed of "tech savvy" people - e.g. computer components suppliers.
3) their site provides an integrated discussion/community area and goods purchasing forum (i.e. online building suppliers with an associated DIY site).
4) they can genuinely offer savings over traditional stores, something which (once you factor tax breaks) only a few online businesses actually achieve.
Some sites fall into several of these categories - for example I would say that Amazon fulfilled almost all of them:
* you can search for specific titles rather than having to physically hunt through the shelves
* the ratings and suggestions systems make it easy to locate new books
* I suspect a significant amount of their business is in computer literature.
I'm a very long way from convinced that an online shop which is just "online" will ever be particularly useful for a small business. Consumer hubs and price comparison services make it very difficult for a small business to actually generate very much interest in their site let alone generate many sales, and what many
Like I said you have to add some value with your site if you want it to make money - and I think that this is what the VC community is just waking up to - of course I'm not a VC so maybe that's just an impression from the media.
Other people in this discussion have talked about tipping sites or selling very low ticket goods, the problem is that presently (in a distributed environment) its not very easy to process those small amounts. Online payments really need to be authorised against something or the fraud risks become enormous - now this is either by using a traditional bank account which incurs minimum authorisation charges (typically around 5p which makes truly micro payments impossible) or against some form of online stored value system (either developed as part of your site or externally i.e. Beanze).
However micropayment solutions are still a long way from pervasive and I'm not sure if consumers are really interested - people like to have a single account that they keep track of and having to occasionally charge a separate online account seems like something of an overhead - perhaps if the account was set up with a direct debit so that your online "charge" was transparently topped up when required it would be easy to use - but then you run the risk of huge bills being generated if the architecture is compromised.
In short I'm not convinced that micropayment based revenue is going to be practical within the next few years - I'd be happy to hear different but I just don't see the social or economic hurdles evaporating.
just thoughts
Tom
depends on what you mean - OpenMarket have a goods type called micropayments in their Transact product. Basically a lump payment made to a server which then handles pay per view URL's on behalf of the merchant (well near enough for the sake of this conversation). I'm 100% sure they patented that system.
However the term micro-payment is usually applied to any transaction below £1 (or $) so I don't think you could patent the ability to process a small amount of money (unless you are a major bank of course *grin*)
actually one of the major problems with micro payments over the net is the overhead of processing these - you either have a choice of not auth'ing the card (in which case you could see fraud rates as high as 70%) or you have a processing cost which could be greater than the ammount in question.
this case isn't about whats good for Slashdotter's or MS - it should be about whats good for consumers and that tends to be healthy competition.
To my mind a more formidable MS would be a good thing.
IMHO
maths grads == great computer scientists
comp sci grads == great computer engineers
I'm comp sci. and was itching to get into industry for the last 2 of my 4 year course - it was all good stuff to learn but I wanted to build systems.
Couldn't see a spec in the artical, but we have a couple here (dual Xeon 500, 1Gb ram, RAID) and their list price is around 16k.
and just to be picky 64x10k=640k or over half a million!
Tom
Personally I want to get one as soon as they come out - I've been wondering about buyind a DVD player for a while and I'm just gonna hold out for the PS2 - that way I get the games console part of it for practially free!
They said it was illegal to export - that doesn't mean that local law overseas prohibits owning it!
I don't belive the problem here is anything to do with moral "should it" "shouldn't it" be taxed discussions - likewise I don't belive that this is legislation which applies solely to web based trade. Its much more to do with the practicality of deciding where the tax liability is incurred and then applying the tax in that location - mostly this problem is down to deciding where the transaction takes place.
e.g. Company based in USA, server hosted in Bermuda, customer in Canada.
Whose tax rules do you use? Currently people seem to be tending towards using the location of the server for any service (i.e. online content) and for the location of the customer for physical purchases. (Well this is the situation in europe anyway). But this is far from clear, it wouldn't suprise me if at some point people start to incure taxes at all sites involved!
Under current systems companies account for their VAT liability - but with a serious increase in cross boarder trade those companies need to be VAT registered in multiple locations if they ship above a certain revenue to those locations.
The only real alternative is to make every individual responsible for handerling thier own VAT (sales tax) liability and applying the tax laws local to the customer - I don't belive that this is going to be practical anytime soon due to shear volume!
Hence, if anybody really wants to tax trade over the net there needs to be a clear descision about where transactions take place - and nobody wants to do that because all pure net businesses will instantly "relocate" for the tax advantages!
Actually I can personally contradict this - I wrote my final year project (in 1998) up on a Mac SE. Now while this is an old machine - it is still perfectly caperble of word processing meaning that I only had to hunt for lab space in order to print.
On the otherhand I have also used both Staroffice and Wordperfect 8 under linux (neither extensively). These also seem to work fine for writing documents for printing elsewhere (never did get my printer properly configured).
But in addition I could have written code and surfed the net (had I not been studiously writting up my project that is)!.
In essence for the average user one WP is much like another, basic style support and some form of graphical import is all thats really essential to the base user.
just thoughts!
Interesting viewpoint - however if I had that number of machines at home I'd probably want a rack mount too!
Just cause somebody wants their kit neat doesn't make them anal... imho properly organised kit is usually easier to access than a rats nest of wires.
As for advocating messy code in in OS enviroments I think you are wrong - clean code is easy to maintain - the advantage of OS is to have multiple people working on a project - if the code is "intricate" then there are fewer people who can join in.
Make stuff simple - make it clean. But of course that is just personal preference.
Tom
COM
From what I understand a breakup is not punishment - its a way of removing the monopoly. From my laymans point of view I would expect punishment to be in the form of either a HUGE fine or of some forced concessions to competitors.
but do they work cheaper?
Its one of the things I liked about my university course... over my four years I covered Ada, PROLOG, Gopher, Z80, 68000, C/C++ and a host of other languages.
But more importantly we covered data structures, alorithm analysis and design looked at software lifecycles and examined lots of specialist areas (real time systems, os design, compiler design, AI's/computer vision, cryptography, etc...)
Nowadays I'm working at an e-commerce consultancy... most of my time is spent integrating with payment gateways some of which are written in Cobol on a VAX and some of which are have CORBA idl mappings.
I'm usually working in languages/enviroments which I haven't specifically seen before the start of the week - but so far I have always seen the underlying concept before.
I guess thats the difference a good 4 year course makes.
Tom
Yup all fair points... I wasn't suggesting this as a certaintly just one possability. More to the point I wasn't suggesting that there was a disproportionate amount of geniuses in those populations... merely that they where not being fully tapped as part of the labour force.
Moreover my main point was simply to agree with Rob that there are likely to be a number of unpredictable shocks to the IT/computing industry.
Given that 3rd world tech shops have been a concern for a while I am perfectly happy to accept that they are unlikely to be that shock... but there are plenty of unknowns left out there.
TomI really don't belive that post is worth a 5 - but whats worrying is that lots of this readership seems to. This is the same sort of self agrandesment as we get with any Linux story - look at this objectivally guys if you don't you run very serious financial risks.
What Rob actually said wasn't that there was going to be a huge increase in the supply of programmers but that there was almost certain to be a structural shock to the industry... he specifically said that he could not predict the source of this shock.
This could be that:
a) More problems are solved by pre-packed solutions - hence less need for custom solutions or from a sys-admin point of view maybe vast leaps forward are going to be made in reliability.
b) A new language comes out which lower quality programmers can use to achive equivalent results.
c) Large amounts of new labour become available - look at all the companies which have experimented with outsourcing their projects to 3rd world techies... these guys are just as bright, work just as hard (or harder) and cost fractions of a western worker.
The likelyhood remains that if this industry (IT et al) where to remain structurally the same then salaries for techies will level off and then fall in the medium term due to increased supplies of new graduates (of which I am one).
But look at the longer term and the chances of some random inovation reducing the numbers of us which are required or much more likely changing the TYPE of techie that is required is pretty good.
We are still an infantile industry - demand is high prices are volotile but can it really last indefinatly.
I doubt it and I'm certainly not going to bet my future/pension and morgage on it.
just thoughts
Tom
A couple of people have suggested mirroring the site... but that sounds like quite a lot of work if you want to expand it accross all stories posted.
/. proxy (storyproxy.slashdot.org)? Would it be possible to set it up so that only URL's (and thier sub directories?) actually posted within main stories would be cached?
/. but I guess it would still cost quite a few $$$ in bandwidth and machines.
/. proxy?
howabout a
It would have a lower maintenance than trying to mirror every site that get posted on
Besides think what it would do to the polls if 30-40% of users all went throught he
*grin*
just thoughts
Tom
I think the comment may have been more directed at people who include all of their major code in GUI event handlers rather than having the handler call another procedure.
If all you do is propogate calls to another procedure (probably in a separate object file) then your maintenance becomes much simpler.