I'm not sure why anyone refers to employment as a games developer as "steady". They are precarious outfits, pathetically dependant upon "hits" that may or may not come again, until they burn you out and drop you like a stone.
An easy explanation for developers "going rogue" is that the pay is so very very bad that the difference between unemployment and salary whilst you write the code is so small that it is not as hard a decision as in other lines of work.
You will also have been made promises by your current employer. These now have no value, indeed reminding them of their commitments, even when made in writing will just get you added to the list of "troublemakers", which will hurt your chances of choosing when you leave.
As well as the "great opportunities in HP" BS, there will be loads of what sound like firm promises of training in new sexy tech, promotion, or whatever they think will convince you to stay for a while.
Also be very aware that outsourcers really do not care if the work you do is critical or hard to replace.
I'm a headhunter, so I genuinely believe job ads. Yes. really. I believe HP job ads that say "Business Development $X+ share options" "Pre Sales Support $X/2 + Package" "Post Sales Support $X/4 + Free Coffee"
Getting the business is what they care about, delivery is simply not an issue. HP are not much worse than any other outsourcer, it is simply the way that sort of business operates.
Although many factors are beyond your control, you can give the impression that you are really positive about the whole thing. Cooperate enthusiastically about the whole process. You can't stop it, or even slow it down, your goal should be to ensure that management think you are "part of the team" and "have the right attitude". You need to do wholly BS stuff like ask those controlling this mess if there is some way you can help. As a techie turned CIO turned headhunter I laugh openly at the "great opportunities" at HP or any outsourcer. My former colleague at PC Magazine Guy Kewney refers to these as "Industry Standard Lies". They have as much credibility as the many fine offers I get in emails from Nigeria and China for wealth and health. But you must not share my laughter. You must sound impressed with these fake offers, maybe even apply for some. This is best done as innocent questions, like asking the new management about them, and how you could apply. They will be selling them to you, so we now have both of you faking it to each other. That will look good when they decide who to dump. You may wonder if they will be taken in by this fake enthusiasm. The odds are better than you think, unless you have already met some HP managers, and then been amazed that people like this are allowed to be in charge of anything. They're not exactly very bright are they ? You will want to leave of course. Anyone who has seen outsourcing knows that the good people will want to walk. But in any market, you want to be the one who decides when you leave. Buy some time and maybe the horse will learn to sing.
You are right. A respectable C++ QD is on around £80K base. The distribution includes developers on £400K including bonus (OK, not *this* year, but still not poor). 200 K is above average, but not unusual. Entry level is 'merely' twice what a games programmer gets.
I headhunt these people, so the numbers I cite are real. The skills are remarkably similar. High end investment banking/hedge fund S/w is written in C++, with bits of maths/physics thrown in.
The working environment is wholly better. Bankers work shorter hours than games developers, and treat their staff with something that approximates to respect, usually
The technologies overlap in ways that are not obvious at all. GPUs and FPGAs have a growing role, and PPUs are being poked at to see what they do. Banks use a mix of open source, closed, and a large amount they knocked up themselves.
My company has a variety of contact lists, and if any of them were to "leak", by CC etc, I'd start getting emails on addresses that *look* like real people but are in fact aliases for me.
If you boss spams like this, there exists the possibility that the other firm have taken this elementary precaution, which may be anything from seriously embarrassing to legally expensive.
Good point. If you do it by surface area, then the British Empire is in fact still truly huge.
Britain asserts sovereignty over a 200 mile radius from any land it controls that isn't controlled by some else.
Thus the Falkands, and the bits of land around there, give about half a million square miles of sea. Ditto Pitcairn etc.
As I recall, The Falklands became a Royal Navy base in part because coal was easily extracted locally. The surveys are not yet incomplete, but there appears to be serious amounts of oil there.
It is at this point that a student of British history will point out that Britain is an ally of the USA, but that the USA is not ally of Britain.
There are moves to make British honours more modern and politcally correct. The British Empire is of course Belize, the Falkland Island, Rockall, bits of Antarctica, and Gibraltar.
Sadly that does mean instead of Commander of the British Empire, one would be Facilitator of the move towards general consensus on climate change. or Chairperson of goodwill towards all nations.
Firstly of course the award is not decided by the Queen. She is a constitutional monarch, and all such decisions are in theory made by minsters.
Even they don't make most of the minor ones, delegating it to committees.
This level of award confers such rights as your daughter being able to marry in St. Pauls Cathedral (the one Princess Di got married in), but little else.
I'd also take exception to the notion that game development in the UK or elsewhere is a desirable career.
It is so badly paid that it cannot be offshored to India because Indians won't work for that little.
EA games and several other firms have been prosecuted for violations of minimum wage laws.
Game developers are treated with a contempt that I have not seen in any industry (I've been a chemist, worked in banking, education, IT, journalism, night clubs and most recently headhunting), and none treat their staff so badly. Even the one nude model I know gets more respect from her employers.
I agree, I see this as a loan, not a sale.
I review software for various publications, and I used to do hardware as well.
S/W firms do not ask for their disks back, but I do not feel I have an ethical right to sell this stuff. Same with h/w.
The nominal value of the s/w I've had over the years is in the 100's of K$
But I would regard it as wholly wrong to bung the 1500-2000 disks I have been sent on Ebay, regardless of the legalities.
I do use it for my own purposes, and the s/w firms like this of course since it means a minor journo is more embraced into their product.
I do know that some DJs make a good % of their income from selling CDs they are sent. Some come in promotional packaging which gives them more value to collectors or serious fans.
To me there is an issue of potential corruption here. I discount the idea that a major label is significantly hurt by a DJ giving a CD to a friend, but a radio DJ is to me a sort of journalist, and if a supplier gives me something that has a significant realisable cash value, then there is a conflict of interest here.
Not one CD of course, but if (say) Umg gives a DJ 50 CDs a week, (they release far more), and he selles them, that's a few hundred bucks per week.
When I was at PC Magazine, one staffer requested to be allowed to remain a music reviewer because of the huge wave of free music she got in a narrow speciality.
Entertainingly, this may even protect DJs a little against the tax authorities. Say 100 CDs a week, that's 5000 CDs, with a face value of $75-100,000.
Is this income ?
Is it declared ?
I absolutely, agree, that was what I was trying to say.
But just because we both think it is a tragically bad idea to buy in most of the tech for CC, does not mean it won't happen.
A related issue is of course that CC's tools must be assembled from things that people have already built, before they joined up. Either by people who've been at CC longer, or that they bring with them.
Creating a rational level of trust (as opposed to ordering people to trust) will be an interesting challenge.
Yes,I agree with your position, but I should have gone deeper into what I mean about "trained"
Beyond a discussion of technical skills, and how to acquire them is the training that separates someone who can shoot straight from a soldier. CC can draw upon the resources of a professional army with centuries of experience in getting people to work in certain ways.
I agree that you represent the mainsteam ethical position here, but I suggest that you do not represent the ideal hire for CC. An armed force cannot work if every single person has to agree with not only the righteousness of the war, but also the technical merit of the style of fighting. It would take too long to do anything. But you don't want Asimovian style robots either.
I'm now retired from full time code writing and become a headhunter (no it's not a pitch for CC business, I'm British so can't play)
CC has a personnel problem unlike the physical military in that they don't have a clear separation between people that prepare the technology, and those that use it against the opposition. Actually that is an assumption I fear I must question. I've worked at the level you have (and occasionally down to hardware), and I think we can both agree that the idea of trusting equipment in the way that an pilot must use black boxes is quite alien to us. Sure you can use an Intel processor, or a hacked OS/2 kernel, but you'd want to know you can poke around anywhere.
I don't know how CC does its procurement, but it's now not a trivial little organisation, so defence contractors will want their share of the pork barrel. Government contractors aren't exactly fans of open source, even with their customers, so I fear that CC will get shafted with grey boxes that meet the 300 page official specification for being "open", but in reality as are open as Microsoft Excel. You then have to call them in, and explain the problem (to people who weren't good enough at C++ for me to shove them into an investment bank for 3 times the money), who then come back 6 months later with the wrong thing
Most black hat hackers are self taught and/or doing it for fun. Currently... They may have been formally taught various bits of programming and networking, but in some respects are a sort of Davy Crockett with a sharp eye rather than a West Point education.
But CC will no doubt be giving its staff a full rounded training, based upon a growing institutional memory, and experts from other parts of the US government, and academics. Being a military outfit, I assume it is configured so that if something awful happens, the organisation still survives, and tries to learn from the setback, even if there are losses. Most of its opponents have a shallow resource base, and actually need to be quite risk averse, since they could not survive a serious problem. CC staff will benefit from this expensively gained experience, and of course often be able to learn things that you could not try if you knew than any error would mean terribly bad personal experiences.
All good news.
But of course, even with careful screening of backgrounds, and various forms of peer review, some will go bad.
These people will be orders of magnitude more dangerous than the random "background noise" hackers. Although sadly some former servicemen go bad this is typically "retail" level damage, often to themselves. Rarely does this get to a level that is beyond local law enforcement, partly because they no longer have access to the infrastructure of the army. A military pilot who once commanded a bomber armed with nuclear weapons is no more dangerous than his civilian counterpart, and so on.
But in cyberwarfare the playing field is much more even. Outside of the 'A Team', the idea that former servicemen could even survive an attack on substantial conventional forces, much less win is plain dumb. So it is a new type of personnel challenge.
But 5 years from now there will be former Cybercommand veterans, complete with a (very discreet) badge and maybe even reunion parties. Mostly their path will be like former pilots, or other specialists who have a ready market in civilian life. But not always...
They will outclass the current generation of hackers, indeed if they did not, then CC would have not have done it's job properly. That to me is a possible issue.
I don't seriously expect an answer in a public forum, but I wonder if plans are yet in place to somehow manage the risk of this, without seriously impacting their utility whilst in the service.
I wrote this piece and I was at pains to say that I was joining the dots to work out how it might have come about. It's based upon many years in big banks, and the kind and anonymous help of people who are trading big numbers in derivatives every day, as well as IT people in banks.
Inevitably unfolding events will show that I joined some wrong dots, and missed some.
Indeed, my focus has drifted away from Excel a little and towards SunGard. Although not the most distinguished of investigative journalists (I'm really a headhunter these days), I smell something bad there. At first I thought that their involvement was peripheral, since if M.Kerviel had the right passwords, their system could not be blamed for any misreporting of the state of SocGen's position.
However their PR people are certainly giving me cause to question an assumption that I now suspect betrays me.
>I imagine that this is largely done due to the constraints of what you can teach in a course.
I teach only unmanaged (currently) though I suppose I'll have to teach that, actually what I will teach is the ugly shit like how do deal with s/w that is a mix of managed and unamanaged code. Already I teach how to write code when you control the "top" and the "bottom", but the middle is code controlled by someone else.
>ese tools would be out of place in an OS but you have to teach them unless your number of students is small and you can hand debug every program that's being written - Actually that is *exactly* how it was handled when I was an undergrad. Alongside the lecturer was some collection of postgrads, and the smarter end of the older students. Cost a bit of money, but if a student put up his hand, he got help, hell we even did that with Pascal. Aside from the money, I liked it because I learned how to deal with other people's bugs. Even now I regard myself as class act in spotting the bugs one gets handed to spot in interviews.
Could I be a bit pedantic, that I am less interested in detailed knowledge, than detailed understanding. You examples are good, and I would add higher level ideas like "how does optimality conflict with fairness", and the way I might identify a superior candidate is "how do you decide the right level". That I shall (rather pretentiously) call "wisdom". Making good choices, and thjat is what I want from CS graduates (or anything else), and is my ambition for those I teach myself.
Must be said that my first reaction is to throw hardware at such problems. Although I seem to have ended up painted as a Java hater, I am more relaxed about it's performance problems than many people. Unless you are doing real time(ish) work more h/w is a sure kill for a fixed budget.
Database issues more often resist hardware solutions, often this is down to locking issues where "merely" doubling CPU buys you very little. But VaR is by it's nature not write-heavy, so locking ought not to be screwing with you, unless the SPs are badly written. Oh, hell tell me that you weren't doing row by row JDBC with locks and cursors ? Please tell me that, I'm giving the final Dominic Connor memorial lecture tomorrow (1/2 a mile from the Angel office), and I need not to lose sleep tonight.
In extremis, for VaR I might even go for detaching the files the DB is on, copying them to another server, and spitting the load that way.
Obviously I don't know your setup well enough to know if my ideas are crap or not, but I think a useful bit of a CS course ought to be that sort of hack. Actually it would make a pretty good O'Reilly book, let's do it right here, rigght now. "Extreme performance hacks". Picture of a chainsaw on the cover, or to be more O'Reilly, a velociraptor.
I agree, and see it as comparable to the difference between economics and business studies, or indeed between physics and engineering. One issue I've faced both as a foot soldier and a CIO is that many firms say "the techies do not know/care enough about the business", yet in my experience >90% of the time the techies are very interested in the business and hungry to learn more. But when they ask for training, or spend time talking to domain experts they get stomped on to "focus" on the technical stuff. Techies can make very good salesmen, OK, not always, but I sold the British Treasury the most expensive thing they ever bought for themselves, and a big thing in this was that I could answer any question on any aspect of the system. Thus when I said it could do X, or that A was better than B, I was believed. Sales people suffer horribly when all their understanding is second hand and way outside their understanding. Sales skills could help a lot, even for purely internal work. Most companies have fallen into a "suck" model of IT requirements. They say "we want it to do X", and they typically get something between W and Z. But as techies we have all discussed cool things they system could be made to do. Most of which are of course useless, or actually illegal or damaging, but I will share with you having sat in meetings with marketing, finance and sales, that their ideas average rather worse. The IT orginiated ideas need to be sold to the rest of the business, not only because they may be useful, but also because they tend to be a lot more achievable since they come from the people who do things to the existing setup. Even when they are crap ideas, there should be feedback, so that they can learn to produce better ideas.
Also bonus and pay do not reflect contribution to the bottom line. Not only is the number too small it is so disconnected with achievement that I will share with you as a headhunter that some techies tell me that they reckon they were overpaid. The bonus oprocess is polluted by heavy noise, and the fact that IT bonuses are set by IT directors who see themselves as part of the "management team", not IT. Thus they have only a vague idea what their people do or contribute. One would never put an accountant in charge of sales, but they frequently end up controlling IT. These are not even good accountants. A good accountant gets to be finance director, or whatever, do you think a top flight accountant *wants* to manage IT. No. So you have 3rd rate beancounters running things.
>I hope you mean $30k-$40k more than the regular salary. I'm a headhunter, the bank pays me 25% of first year salary, hence those numbers at entry level. It's hard to say all that precisely what it's worth, since in many cases it does you no good at all if your work does not intersect with that type of investment banking, whilst writing I see a curve, basically your chances of making seriously good money at least double, and sets the bounds of what you might reasonably expect to go up. I suppose you model it as a large box of lottery tickets.
You're right few undergrads have got that far for the top jobs, but SP helps be a member of these teams. A QD with SP can get pretty decent money, and that can be got from some ugrad degrees. That's actually pretty rare though since most at that level do further study.
We do look at EE, and it is a better hunting ground than CS, which is not a criticism of CS, though I do like to see some information theory in the skills set of a CS, partly for the SP niche, but mostly because I believe that it is a critical underpinning to really "getting" software. It won't help you do loops, or solve pointer bugs, but as an enabler of high level thinking about systems I cannot imagine how you can think properly without it. The fact that it is useful in data compression and SP, is to me a lucky coincidence, not the real motivation.
I don't mind slow disk I/O, for the sort of work I pimp for, it is typically not a major issue for the trading end. It is possible for an entity to be using these without me knowing, but I'd bet money it's a first cut/pathfinding deployment, else my deep radar might well have picked it up. I forget the numbers, but I recall liking the heat dissipation, on the bleeding edge of finance we are already fretting over heat, thus the using of performance is joule instructions per second. This is a marvellously deep recurrent pattern. It turns out that best way of modelling derivatives is based upon thermodynamics starting off with the diffusion you get in a simple heated metal bar to convection/advection. Of course these derivatives include options on the energy being consumed to price them, whose dissipation is causing deep stress to some IT managers.
I do think students need to spend some time learning how to understand big complex systems. An O/S is a good example of this, but of course not unique.
Obviously there should be thorough teaching in how to avoid bugs, but I feel that the subject of debugging is neglected, some people graduate with no formal teaching in this important discipline at all. Again you need multiple languages to get there, and some bug types are mutually exclusive. You don't get a lot of pointer issues in Javsa, but C++ does not present you with the issue of not knowing when your freed resources will actually be released.
I don't expect someone to write a whole O/S as undergrad, indeed my perception is that the importance of O/S writing as an activity has been declining for at least a decade. I just like to see people who have worked at as many levels as possible. OS/s can be written in pretty much anything, but C/C++ is by far the dominant language, so if a student is going to take the lid off, that's the tool he should have some skill with. "taking the lid off" is an indicator of the intellectual curiosity that I greatly prize. Elsewhere in this thread people are talking about Apples as stable platforms. I like stability, like any other user, but something feels wrong about a CS who prizes this above all. Fair enough if you use a PC for writing essays, but to me a CS ought to be by instinct someone who cracks open his PC within days of it's delivery.
I refer to myself here as a "user", and that has a passive ring to it, to me a CS should be someone who does things to computers, not someone that computers do things to.
Sorry about the formatting, so much for me being a techie:(
My values as a former manager of techies are similar to yours, but one can hire what is out there, and sometimes the right technical skills do not come in a friendly package. I quite literally busted two buttons off my shirt in the frustration in not being able to kill one techie where he stood, so I have complete sympathy with the need to upgrade people skills. (are you reading this David ?)
My niche is lighter on this requirement, partly because most quants don't have to talk to customers much. As someone who sells such people to banks, I know that the one with better "technical" skills will get the job 9/10, so I serve my customers what they want to buy. But a few have got vetoed because of extreme personality issues, and a few dig themselves a whole by attacking their current employer too violently.
If you will indluge me, I will share part of our Guide, where we echo some of the issues that arise when we ask people why they are leaving their job. For confidentiality I have changed a few nouns, but the sentiments are real.
Why your boss is a jerk If you're leaving your job, it's of course likely to be because you've stopped enjoying the work, or that your job doesn't offer you the opportunities you want. Money may well be a factor and as HH's we don't have a problem hearing this. Money has many problems but at least it is relatively simple as a motivation.
Sometimes however we hear from a candidate that it is because his current boss is a shambling moron whose personality is an unstable mix of dishonesty and ignorance barely held together by malicious greed. His management style draws upon both forms of Marxism, both Groucho and Karl. He can recite "The Art of War" from memory and he frequently quotes from it at meetings (in the original Chinese of course). You feel you have to leave now or you and he will settle your disputes with knives. The IT at your department looks like it's run by EDS, the management are in league with Al Qaeda, compliance has been infiltrated by Accenture and Jack Bauer has told you that the back office wants you dead. Today you found a live rat in your coffee.
>Your case is likely colored by the kind of work that needs to be done in investment banking (and insurance). Yep, an assumption I have tried to make as clear as possible.
>On the other hand, most day-to-day administrative/financial work outside of investment banking is database work (SAP, anyone?), which would require a totally different kind of profile Agreed, and before you ask, I don't find the database theory on many CS courses that impressive either.
I agree about SAP et al, and I'm big on "portable" skills. These tend to be deeper and more theoretical, allowing you to move on when SAP (or VB or Java or PHP) are no long er in demand.
I also agree about college boards, been on one myself. Hanging out with bankers has led me to offer one idea on how colleges should be funded. That is by instead of charging fees, they get a % of your lifetime earnings, say 1 or 2%, collected through the tax system. Although they would care about your first job, this would represent only 5-10% of their rakeoff, so they would have a profound incentive to not only give you long lasting skills, but also to provide alumni learning. The private sector course I teach on (Wilmott CQF) is a lifetime deal, you get to sit in on new lectures to keep up to date, and all are videoed so that you can learn at home or distance. It's cool, but not exactly rocket science, yet I know of no mainstream academic institution that does this.
2% of pay is usually not something most people would go to great lengths to dodge, so they'd get most of it.
I'm older than most people around here, and have had to upgrade my skills more than once, and have made occasional bad bets on what was the next big thing. And "bet" is the right word. When you learn SAP or Java, you are betting that in the next few years there will be a good market for that skill.
Although my posts here have caused some to think of me as dumb, and my current role as a headhunter does not exactly counter that, I am a keen student of trends, and through various channels have better information than most, so my screwups are more my fault than those of others.
I have seen Powerbuilder, OS/2, VB, network admin, Clipper, Nortel Routers, and any number of other skills go from red hot to either death, or mediocre pay with limited career options. I've managed to stay an employed techie longer than most of my peers because I had a first rate grounding, and because I've invested heavily in learning new shit. And yes, I am a techie, companies give me money to make things happen that they can't do themselves, and could get a real job by Monday, tomorrow if wanted to accept an average wage, indeed on my desk is a role working on the next Harry Potter film (as an incredibly minor techie) that I could do, if the pimping went pear shaped.
Are you a Windows fanboy trying to make Apple fanboys look bad ?
Are you really saying Apples don't get virii ?
>I could not give a damn about the way apple treats developers, as long as I can develop Java code on my macs, I don't care. And you aren't aware of any of the recent history on this ?
Actually I don't care any more than you do about Apple developers, but my point was that if someone is going to call Apple the good guys they ought to know their track record.
That's a rational view on Apple, that's what I look for, am more impressed by a smart contrary view than a blind love of Linux, Windows or even Os/2. To me you don't understand something if you can't poke holes in it, even if you agree with the general position.
>And I still don't understand your push on signal processing. You are looking for people making business software for a bank. The short version is that each good entry level SP whizz is worth $30-40K to me:)
>And I still don't understand your push on signal processing. You are looking for people making business software for a bank... By "business s/w" I guess you mean the standard accounting, sales, customer management and inventory tracking type of thing ? We don't do that either.
Short version of why this is useful: Prices of financial things move so randomly that they are often modelled as completely random. This allows you to use the well established maths of stuff like heat diffusion and stuff like Brownian motion to price derivatives. You can have long arguments about with random distribution to use, but most work uses the bog standard normal distribution.
But we do observe that that some things are "caused" or at least correlated to other things. But we are in an environment not unlike SETI, where they may or may not a signal but with appalling signal to noise ratios. A priori we don't even know what signal we are looking for.
The media likes to report that changes in interest rates "cause" stock prices to go up or down, which is sort of true. But if you look at the raw data, it is not always that easy. Indeed an interest rate cut can "cause" share prices to go down if the market assumed the cut was coming, or if it was expecting q bigger one. Sometimes another factor swamps this signal anyway. Also this is a very obvious signal so it's hard to beat others to exploit it, so the search is for other signals that imply changes either in the absolute level of price, or more recently in the volatility of the price. Exploiting this is a good earner for the investment banks and hedge funds, when they get it right.
Depending on the market the signals come in from about 100 to 0.01 Hz in the form of prices and trades. For options on indexes the data flow can be 10KHz or more. Finding a pattern in that is tough, as is not mistaking coincidence for signal.
This is a speciality, I'd guess around 0.1% of IT people in banks use any significant amount of SP, but that fraction do make good money.
Most CS courses do not seem to have SP, it typically being in EE, though at some places students may pick from both departments, and there are many hybrids.
Slashdot Mirror
I'm not sure why anyone refers to employment as a games developer as "steady". They are precarious outfits, pathetically dependant upon "hits" that may or may not come again, until they burn you out and drop you like a stone.
An easy explanation for developers "going rogue" is that the pay is so very very bad that the difference between unemployment and salary whilst you write the code is so small that it is not as hard a decision as in other lines of work.
You will also have been made promises by your current employer.
These now have no value, indeed reminding them of their commitments, even when made in writing will just get you added to the list of "troublemakers", which will hurt your chances of choosing when you leave.
As well as the "great opportunities in HP" BS, there will be loads of what sound like firm promises of training in new sexy tech, promotion, or whatever they think will convince you to stay for a while.
Also be very aware that outsourcers really do not care if the work you do is critical or hard to replace.
I'm a headhunter, so I genuinely believe job ads.
Yes. really.
I believe HP job ads that say
"Business Development $X+ share options"
"Pre Sales Support $X/2 + Package"
"Post Sales Support $X/4 + Free Coffee"
Getting the business is what they care about, delivery is simply not an issue. HP are not much worse than any other outsourcer, it is simply the way that sort of business operates.
Although many factors are beyond your control, you can give the impression that you are really positive about the whole thing. Cooperate enthusiastically about the whole process. You can't stop it, or even slow it down, your goal should be to ensure that management think you are "part of the team" and "have the right attitude".
You need to do wholly BS stuff like ask those controlling this mess if there is some way you can help.
As a techie turned CIO turned headhunter I laugh openly at the "great opportunities" at HP or any outsourcer. My former colleague at PC Magazine Guy Kewney refers to these as "Industry Standard Lies".
They have as much credibility as the many fine offers I get in emails from Nigeria and China for wealth and health.
But you must not share my laughter.
You must sound impressed with these fake offers, maybe even apply for some. This is best done as innocent questions, like asking the new management about them, and how you could apply. They will be selling them to you, so we now have both of you faking it to each other. That will look good when they decide who to dump.
You may wonder if they will be taken in by this fake enthusiasm. The odds are better than you think, unless you have already met some HP managers, and then been amazed that people like this are allowed to be in charge of anything.
They're not exactly very bright are they ?
You will want to leave of course. Anyone who has seen outsourcing knows that the good people will want to walk.
But in any market, you want to be the one who decides when you leave. Buy some time and maybe the horse will learn to sing.
You are right.
A respectable C++ QD is on around £80K base.
The distribution includes developers on £400K including bonus (OK, not *this* year, but still not poor). 200 K is above average, but not unusual.
Entry level is 'merely' twice what a games programmer gets.
I headhunt these people, so the numbers I cite are real.
The skills are remarkably similar. High end investment banking/hedge fund S/w is written in C++, with bits of maths/physics thrown in.
The working environment is wholly better. Bankers work shorter hours than games developers, and treat their staff with something that approximates to respect, usually
The technologies overlap in ways that are not obvious at all. GPUs and FPGAs have a growing role, and PPUs are being poked at to see what they do.
Banks use a mix of open source, closed, and a large amount they knocked up themselves.
http://www.wilmott.com/blogs/dcfc
My company has a variety of contact lists, and if any of them were to "leak", by CC etc, I'd start getting emails on addresses that *look* like real people but are in fact aliases for me.
If you boss spams like this, there exists the possibility that the other firm have taken this elementary precaution, which may be anything from seriously embarrassing to legally expensive.
Good point.
If you do it by surface area, then the British Empire is in fact still truly huge.
Britain asserts sovereignty over a 200 mile radius from any land it controls that isn't controlled by some else.
Thus the Falkands, and the bits of land around there, give about half a million square miles of sea.
Ditto Pitcairn etc.
As I recall, The Falklands became a Royal Navy base in part because coal was easily extracted locally. The surveys are not yet incomplete, but there appears to be serious amounts of oil there.
It is at this point that a student of British history will point out that Britain is an ally of the USA, but that the USA is not ally of Britain.
There are moves to make British honours more modern and politcally correct. The British Empire is of course Belize, the Falkland Island, Rockall, bits of Antarctica, and Gibraltar.
Sadly that does mean instead of Commander of the British Empire, one would be
Facilitator of the move towards general consensus on climate change.
or
Chairperson of goodwill towards all nations.
Firstly of course the award is not decided by the Queen. She is a constitutional monarch, and all such decisions are in theory made by minsters. Even they don't make most of the minor ones, delegating it to committees. This level of award confers such rights as your daughter being able to marry in St. Pauls Cathedral (the one Princess Di got married in), but little else. I'd also take exception to the notion that game development in the UK or elsewhere is a desirable career. It is so badly paid that it cannot be offshored to India because Indians won't work for that little. EA games and several other firms have been prosecuted for violations of minimum wage laws. Game developers are treated with a contempt that I have not seen in any industry (I've been a chemist, worked in banking, education, IT, journalism, night clubs and most recently headhunting), and none treat their staff so badly. Even the one nude model I know gets more respect from her employers.
I agree, I see this as a loan, not a sale. I review software for various publications, and I used to do hardware as well. S/W firms do not ask for their disks back, but I do not feel I have an ethical right to sell this stuff. Same with h/w. The nominal value of the s/w I've had over the years is in the 100's of K$ But I would regard it as wholly wrong to bung the 1500-2000 disks I have been sent on Ebay, regardless of the legalities. I do use it for my own purposes, and the s/w firms like this of course since it means a minor journo is more embraced into their product. I do know that some DJs make a good % of their income from selling CDs they are sent. Some come in promotional packaging which gives them more value to collectors or serious fans. To me there is an issue of potential corruption here. I discount the idea that a major label is significantly hurt by a DJ giving a CD to a friend, but a radio DJ is to me a sort of journalist, and if a supplier gives me something that has a significant realisable cash value, then there is a conflict of interest here. Not one CD of course, but if (say) Umg gives a DJ 50 CDs a week, (they release far more), and he selles them, that's a few hundred bucks per week. When I was at PC Magazine, one staffer requested to be allowed to remain a music reviewer because of the huge wave of free music she got in a narrow speciality. Entertainingly, this may even protect DJs a little against the tax authorities. Say 100 CDs a week, that's 5000 CDs, with a face value of $75-100,000. Is this income ? Is it declared ?
I absolutely, agree, that was what I was trying to say.
But just because we both think it is a tragically bad idea to buy in most of the tech for CC, does not mean it won't happen.
A related issue is of course that CC's tools must be assembled from things that people have already built, before they joined up.
Either by people who've been at CC longer, or that they bring with them.
Creating a rational level of trust (as opposed to ordering people to trust) will be an interesting challenge.
Yes,I agree with your position, but I should have gone deeper into what I mean about "trained"
Beyond a discussion of technical skills, and how to acquire them is the training that separates someone who can shoot straight from a soldier.
CC can draw upon the resources of a professional army with centuries of experience in getting people to work in certain ways.
I agree that you represent the mainsteam ethical position here, but I suggest that you do not represent the ideal hire for CC.
An armed force cannot work if every single person has to agree with not only the righteousness of the war, but also the technical merit of the style of fighting.
It would take too long to do anything.
But you don't want Asimovian style robots either.
I'm now retired from full time code writing and become a headhunter (no it's not a pitch for CC business, I'm British so can't play)
CC has a personnel problem unlike the physical military in that they don't have a clear separation between people that prepare the technology, and those that use it against the opposition. Actually that is an assumption I fear I must question.
I've worked at the level you have (and occasionally down to hardware), and I think we can both agree that the idea of trusting equipment in the way that an pilot must use black boxes is quite alien to us. Sure you can use an Intel processor, or a hacked OS/2 kernel, but you'd want to know you can poke around anywhere.
I don't know how CC does its procurement, but it's now not a trivial little organisation, so defence contractors will want their share of the pork barrel.
Government contractors aren't exactly fans of open source, even with their customers, so I fear that CC will get shafted with grey boxes that meet the 300 page official specification for being "open", but in reality as are open as Microsoft Excel. You then have to call them in, and explain the problem (to people who weren't good enough at C++ for me to shove them into an investment bank for 3 times the money), who then come back 6 months later with the wrong thing
Most black hat hackers are self taught and/or doing it for fun.
Currently...
They may have been formally taught various bits of programming and networking, but in some respects are a sort of Davy Crockett with a sharp eye rather than a West Point education.
But CC will no doubt be giving its staff a full rounded training, based upon a growing institutional memory, and experts from other parts of the US government, and academics.
Being a military outfit, I assume it is configured so that if something awful happens, the organisation still survives, and tries to learn from the setback, even if there are losses.
Most of its opponents have a shallow resource base, and actually need to be quite risk averse, since they could not survive a serious problem. CC staff will benefit from this expensively gained experience, and of course often be able to learn things that you could not try if you knew than any error would mean terribly bad personal experiences.
All good news.
But of course, even with careful screening of backgrounds, and various forms of peer review, some will go bad.
These people will be orders of magnitude more dangerous than the random "background noise" hackers.
Although sadly some former servicemen go bad this is typically "retail" level damage, often to themselves. Rarely does this get to a level that is beyond local law enforcement, partly because they no longer have access to the infrastructure of the army. A military pilot who once commanded a bomber armed with nuclear weapons is no more dangerous than his civilian counterpart, and so on.
But in cyberwarfare the playing field is much more even. Outside of the 'A Team', the idea that former servicemen could even survive an attack on substantial conventional forces, much less win is plain dumb. So it is a new type of personnel challenge.
But 5 years from now there will be former Cybercommand veterans, complete with a (very discreet) badge and maybe even reunion parties. Mostly their path will be like former pilots, or other specialists who have a ready market in civilian life. But not always...
They will outclass the current generation of hackers, indeed if they did not, then CC would have not have done it's job properly. That to me is a possible issue.
I don't seriously expect an answer in a public forum, but I wonder if plans are yet in place to somehow manage the risk of this, without seriously impacting their utility whilst in the service.
I wrote this piece and I was at pains to say that I was joining the dots to work out how it might have come about.
It's based upon many years in big banks, and the kind and anonymous help of people who are trading big numbers in derivatives every day, as well as IT people in banks.
Inevitably unfolding events will show that I joined some wrong dots, and missed some.
Indeed, my focus has drifted away from Excel a little and towards SunGard. Although not the most distinguished of investigative journalists (I'm really a headhunter these days), I smell something bad there.
At first I thought that their involvement was peripheral, since if M.Kerviel had the right passwords, their system could not be blamed for any misreporting of the state of SocGen's position.
However their PR people are certainly giving me cause to question an assumption that I now suspect betrays me.
>I imagine that this is largely done due to the constraints of what you can teach in a course.
I teach only unmanaged (currently) though I suppose I'll have to teach that, actually what I will teach is the ugly shit like how do deal with s/w that is a mix of managed and unamanaged code. Already I teach how to write code when you control the "top" and the "bottom", but the middle is code controlled by someone else.
>ese tools would be out of place in an OS but you have to teach them unless your number of students is small and you can hand debug every program that's being written -
Actually that is *exactly* how it was handled when I was an undergrad. Alongside the lecturer was some collection of postgrads, and the smarter end of the older students.
Cost a bit of money, but if a student put up his hand, he got help, hell we even did that with Pascal. Aside from the money, I liked it because I learned how to deal with other people's bugs. Even now I regard myself as class act in spotting the bugs one gets handed to spot in interviews.
Could I be a bit pedantic, that I am less interested in detailed knowledge, than detailed understanding.
You examples are good, and I would add higher level ideas like "how does optimality conflict with fairness", and the way I might identify a superior candidate is "how do you decide the right level". That I shall (rather pretentiously) call "wisdom". Making good choices, and thjat is what I want from CS graduates (or anything else), and is my ambition for those I teach myself.
Must be said that my first reaction is to throw hardware at such problems.
Although I seem to have ended up painted as a Java hater, I am more relaxed about it's performance problems than many people.
Unless you are doing real time(ish) work more h/w is a sure kill for a fixed budget.
Database issues more often resist hardware solutions, often this is down to locking issues where "merely" doubling CPU buys you very little.
But VaR is by it's nature not write-heavy, so locking ought not to be screwing with you, unless the SPs are badly written.
Oh, hell tell me that you weren't doing row by row JDBC with locks and cursors ?
Please tell me that, I'm giving the final Dominic Connor memorial lecture tomorrow (1/2 a mile from the Angel office), and I need not to lose sleep tonight.
In extremis, for VaR I might even go for detaching the files the DB is on, copying them to another server, and spitting the load that way.
Obviously I don't know your setup well enough to know if my ideas are crap or not, but I think a useful bit of a CS course ought to be that sort of hack.
Actually it would make a pretty good O'Reilly book, let's do it right here, rigght now.
"Extreme performance hacks".
Picture of a chainsaw on the cover, or to be more O'Reilly, a velociraptor.
I agree, and see it as comparable to the difference between economics and business studies, or indeed between physics and engineering.
One issue I've faced both as a foot soldier and a CIO is that many firms say "the techies do not know/care enough about the business", yet in my experience >90% of the time the techies are very interested in the business and hungry to learn more. But when they ask for training, or spend time talking to domain experts they get stomped on to "focus" on the technical stuff.
Techies can make very good salesmen, OK, not always, but I sold the British Treasury the most expensive thing they ever bought for themselves, and a big thing in this was that I could answer any question on any aspect of the system. Thus when I said it could do X, or that A was better than B, I was believed. Sales people suffer horribly when all their understanding is second hand and way outside their understanding.
Sales skills could help a lot, even for purely internal work. Most companies have fallen into a "suck" model of IT requirements. They say "we want it to do X", and they typically get something between W and Z. But as techies we have all discussed cool things they system could be made to do. Most of which are of course useless, or actually illegal or damaging, but I will share with you having sat in meetings with marketing, finance and sales, that their ideas average rather worse. The IT orginiated ideas need to be sold to the rest of the business, not only because they may be useful, but also because they tend to be a lot more achievable since they come from the people who do things to the existing setup.
Even when they are crap ideas, there should be feedback, so that they can learn to produce better ideas.
Also bonus and pay do not reflect contribution to the bottom line. Not only is the number too small it is so disconnected with achievement that I will share with you as a headhunter that some techies tell me that they reckon they were overpaid. The bonus oprocess is polluted by heavy noise, and the fact that IT bonuses are set by IT directors who see themselves as part of the "management team", not IT. Thus they have only a vague idea what their people do or contribute. One would never put an accountant in charge of sales, but they frequently end up controlling IT. These are not even good accountants. A good accountant gets to be finance director, or whatever, do you think a top flight accountant *wants* to manage IT. No. So you have 3rd rate beancounters running things.
>I hope you mean $30k-$40k more than the regular salary.
I'm a headhunter, the bank pays me 25% of first year salary, hence those numbers at entry level.
It's hard to say all that precisely what it's worth, since in many cases it does you no good at all if your work does not intersect with that type of investment banking, whilst writing I see a curve, basically your chances of making seriously good money at least double, and sets the bounds of what you might reasonably expect to go up.
I suppose you model it as a large box of lottery tickets.
You're right few undergrads have got that far for the top jobs, but SP helps be a member of these teams. A QD with SP can get pretty decent money, and that can be got from some ugrad degrees. That's actually pretty rare though since most at that level do further study.
We do look at EE, and it is a better hunting ground than CS, which is not a criticism of CS, though I do like to see some information theory in the skills set of a CS, partly for the SP niche, but mostly because I believe that it is a critical underpinning to really "getting" software. It won't help you do loops, or solve pointer bugs, but as an enabler of high level thinking about systems I cannot imagine how you can think properly without it. The fact that it is useful in data compression and SP, is to me a lucky coincidence, not the real motivation.
I don't mind slow disk I/O, for the sort of work I pimp for, it is typically not a major issue for the trading end.
It is possible for an entity to be using these without me knowing, but I'd bet money it's a first cut/pathfinding deployment, else my deep radar might well have picked it up.
I forget the numbers, but I recall liking the heat dissipation, on the bleeding edge of finance we are already fretting over heat, thus the using of performance is joule instructions per second. This is a marvellously deep recurrent pattern. It turns out that best way of modelling derivatives is based upon thermodynamics starting off with the diffusion you get in a simple heated metal bar to convection/advection. Of course these derivatives include options on the energy being consumed to price them, whose dissipation is causing deep stress to some IT managers.
I do think students need to spend some time learning how to understand big complex systems.
An O/S is a good example of this, but of course not unique.
Obviously there should be thorough teaching in how to avoid bugs, but I feel that the subject of debugging is neglected, some people graduate with no formal teaching in this important discipline at all. Again you need multiple languages to get there, and some bug types are mutually exclusive. You don't get a lot of pointer issues in Javsa, but C++ does not present you with the issue of not knowing when your freed resources will actually be released.
I don't expect someone to write a whole O/S as undergrad, indeed my perception is that the importance of O/S writing as an activity has been declining for at least a decade. I just like to see people who have worked at as many levels as possible.
OS/s can be written in pretty much anything, but C/C++ is by far the dominant language, so if a student is going to take the lid off, that's the tool he should have some skill with.
"taking the lid off" is an indicator of the intellectual curiosity that I greatly prize. Elsewhere in this thread people are talking about Apples as stable platforms.
I like stability, like any other user, but something feels wrong about a CS who prizes this above all. Fair enough if you use a PC for writing essays, but to me a CS ought to be by instinct someone who cracks open his PC within days of it's delivery.
I refer to myself here as a "user", and that has a passive ring to it, to me a CS should be someone who does things to computers, not someone that computers do things to.
I apologise for wasting the time of someone who is obviously a better man than I.
Sorry about the formatting, so much for me being a techie :(
My values as a former manager of techies are similar to yours, but one can hire what is out there, and sometimes the right technical skills do not come in a friendly package.
I quite literally busted two buttons off my shirt in the frustration in not being able to kill one techie where he stood, so I have complete sympathy with the need to upgrade people skills. (are you reading this David ?)
My niche is lighter on this requirement, partly because most quants don't have to talk to customers much. As someone who sells such people to banks, I know that the one with better "technical" skills will get the job 9/10, so I serve my customers what they want to buy.
But a few have got vetoed because of extreme personality issues, and a few dig themselves a whole by attacking their current employer too violently.
If you will indluge me, I will share part of our Guide, where we echo some of the issues that arise when we ask people why they are leaving their job. For confidentiality I have changed a few nouns, but the sentiments are real.
Why your boss is a jerk
If you're leaving your job, it's of course likely to be because you've stopped enjoying the work, or that your job doesn't offer you the opportunities you want. Money may well be a factor and as HH's we don't have a problem hearing this. Money has many problems but at least it is relatively simple as a motivation.
Sometimes however we hear from a candidate that it is because his current boss is a shambling moron whose personality is an unstable mix of dishonesty and ignorance barely held together by malicious greed. His management style draws upon both forms of Marxism, both Groucho and Karl. He can recite "The Art of War" from memory and he frequently quotes from it at meetings (in the original Chinese of course). You feel you have to leave now or you and he will settle your disputes with knives. The IT at your department looks like it's run by EDS, the management are in league with Al Qaeda, compliance has been infiltrated by Accenture and Jack Bauer has told you that the back office wants you dead. Today you found a live rat in your coffee.
>Your case is likely colored by the kind of work that needs to be done in investment banking (and insurance).
Yep, an assumption I have tried to make as clear as possible.
>On the other hand, most day-to-day administrative/financial work outside of investment banking is database work (SAP, anyone?), which would require a totally different kind of profile
Agreed, and before you ask, I don't find the database theory on many CS courses that impressive either.
I agree about SAP et al, and I'm big on "portable" skills. These tend to be deeper and more theoretical, allowing you to move on when SAP (or VB or Java or PHP) are no long er in demand.
I also agree about college boards, been on one myself. Hanging out with bankers has led me to offer one idea on how colleges should be funded. That is by instead of charging fees, they get a % of your lifetime earnings, say 1 or 2%, collected through the tax system.
Although they would care about your first job, this would represent only 5-10% of their rakeoff, so they would have a profound incentive to not only give you long lasting skills, but also to provide alumni learning. The private sector course I teach on (Wilmott CQF) is a lifetime deal, you get to sit in on new lectures to keep up to date, and all are videoed so that you can learn at home or distance. It's cool, but not exactly rocket science, yet I know of no mainstream academic institution that does this.
2% of pay is usually not something most people would go to great lengths to dodge, so they'd get most of it.
I'm older than most people around here, and have had to upgrade my skills more than once, and have made occasional bad bets on what was the next big thing.
And "bet" is the right word.
When you learn SAP or Java, you are betting that in the next few years there will be a good market for that skill.
Although my posts here have caused some to think of me as dumb, and my current role as a headhunter does not exactly counter that, I am a keen student of trends, and through various channels have better information than most, so my screwups are more my fault than those of others.
I have seen Powerbuilder, OS/2, VB, network admin, Clipper, Nortel Routers, and any number of other skills go from red hot to either death, or mediocre pay with limited career options. I've managed to stay an employed techie longer than most of my peers because I had a first rate grounding, and because I've invested heavily in learning new shit.
And yes, I am a techie, companies give me money to make things happen that they can't do themselves, and could get a real job by Monday, tomorrow if wanted to accept an average wage, indeed on my desk is a role working on the next Harry Potter film (as an incredibly minor techie) that I could do, if the pimping went pear shaped.
Are you a Windows fanboy trying to make Apple fanboys look bad ?
Are you really saying Apples don't get virii ?
>I could not give a damn about the way apple treats developers, as long as I can develop Java code on my macs, I don't care.
And you aren't aware of any of the recent history on this ?
Actually I don't care any more than you do about Apple developers, but my point was that if someone is going to call Apple the good guys they ought to know their track record.
That's a rational view on Apple, that's what I look for, am more impressed by a smart contrary view than a blind love of Linux, Windows or even Os/2.
:)
To me you don't understand something if you can't poke holes in it, even if you agree with the general position.
>And I still don't understand your push on signal processing. You are looking for people making business software for a bank.
The short version is that each good entry level SP whizz is worth $30-40K to me
>And I still don't understand your push on signal processing. You are looking for people making business software for a bank...
By "business s/w" I guess you mean the standard accounting, sales, customer management and inventory tracking type of thing ?
We don't do that either.
Short version of why this is useful:
Prices of financial things move so randomly that they are often modelled as completely random. This allows you to use the well established maths of stuff like heat diffusion and stuff like Brownian motion to price derivatives. You can have long arguments about with random distribution to use, but most work uses the bog standard normal distribution.
But we do observe that that some things are "caused" or at least correlated to other things. But we are in an environment not unlike SETI, where they may or may not a signal but with appalling signal to noise ratios. A priori we don't even know what signal we are looking for.
The media likes to report that changes in interest rates "cause" stock prices to go up or down, which is sort of true. But if you look at the raw data, it is not always that easy. Indeed an interest rate cut can "cause" share prices to go down if the market assumed the cut was coming, or if it was expecting q bigger one.
Sometimes another factor swamps this signal anyway.
Also this is a very obvious signal so it's hard to beat others to exploit it, so the search is for other signals that imply changes either in the absolute level of price, or more recently in the volatility of the price.
Exploiting this is a good earner for the investment banks and hedge funds, when they get it right.
Depending on the market the signals come in from about 100 to 0.01 Hz in the form of prices and trades. For options on indexes the data flow can be 10KHz or more.
Finding a pattern in that is tough, as is not mistaking coincidence for signal.
This is a speciality, I'd guess around 0.1% of IT people in banks use any significant amount of SP, but that fraction do make good money.
Most CS courses do not seem to have SP, it typically being in EE, though at some places students may pick from both departments, and there are many hybrids.