[A portion of this essay appeared in the October 2008 issue of Information Security magazine. My thanks to Dave Farber for a conversation that spurred me to post this expanded version.]

[Small typos corrected in April 2010.]

I'd like to repeat (portions of) a theme I have been speaking about for over a decade. I'll start by taking a long view of computing.

Fifty years ago, IBM introduced the first all-transistor computer (the 7000 series). Transistors were approximately $60 apiece (in current dollars). Secondary storage was about 10 cents per byte (also in current dollars) and had a density of approximately 2000 bits per cubic inch. According to Wikipedia, a working IBM 7090 system with a full 32K of memory (the capacity of the machine) cost about $3,000,000 to purchase -- over $21,000,000 in current dollars. Software, peripherals, and maintenance all cost more. Rental of a system (maintenance included) could be well over $500,000 per month (in 1958 dollars). Other vendors soon brought their own transistorized systems to market, at similar costs.

These early computing systems came without an operating system. However, the costs of having such a system sit idle between jobs (and during I/O) led the field to develop operating systems that supported sharing of hardware to maximize utilization. It also led to the development of user accounts for cost accounting. And all of these soon led to development of security features to ensure that the sharing didn't go too far, and that accounting was not disabled or corrupted. As the hardware evolved and became more capable, the software also evolved and took on new features.

Costs and capabilities of computing hardware have changed by a factor of tens of millions in five decades. Currently, transistors cost less than 1/7800 of a cent apiece in modern CPU chips (Intel Itanium). Assuming I didn't drop a decimal place, that is a drop in price by 7 orders of magnitude.  Ed Lazowska made a presentation a few years ago where he indicated that the number of grains of rice harvested worldwide in 2004 was ten quintillion -- 10 raised to the 18th power. But in 2004, there were also ten quintillion transistors manufactured, and that number has increased faster than the rice harvest ever since. We have more transistors being produced and fielded each year than all the grains of rice harvested in all the countries of the world. Isn't that amazing?

Storage also changed drastically. We have gone from core memory to semiconductor memory. And in secondary storage we have gone from drum memory to disks to SSDs. If we look at consumer disk storage, it is now common to get storage density of better than 500Gb per cubic inch at a cost of less than $.20 per Gb (including enclosure and controller) -- a price drop of nearly 8 orders of magnitude. Of course, weight, size, speed, noise, heat, power, and other factors have all also undergone major changes. To think of it another way, that same presentation by Professor Lazowska, noted that the computerized greeting cards you can buy at the store to record and play back a message to music have more computing power and memory in them than some of those multi-million $ computers of the 1950s, all for under $10.

Yet, despite these incredible transformations, the operating systems, databases, languages, and more that we use are still basically the designs we came up with in the 1960s to make the best use of limited, expensive, shared equipment. More to the theme of this blog, overall information security is almost certainly worse now than it was in the 1960s. We're still suffering from problems known for decades, and systems are still being built with intrinsic weaknesses, yet now we have more to lose with more valuable information coming online every week.

Why have we failed to make appreciable progress with the software? In part, it is because we've been busy trying to advance on every front. Partially, it is because it is simpler to replace the underlying hardware with something faster, thus getting a visible performance gain. This helps mask the ongoing lack of quality and progression to really new ideas. As well, the speed with which the field of computing (development and application) moves is incredible, and few have the time or inclination to step back and re-examine first principles. This includes old habits such as the sense of importance in making code "small" even to the point of leaving out internal consistency checks and error handling. (Y2K was not a one-time fluke -- it's an instance of an institutional bad habit.)

Another such habit is that of trying to build every system to have the capability to perform every task. There is a general lack of awareness that security needs are different for different applications and environments; instead, people seek uniformity of OS, hardware architecture, programming languages and beyond, all with maximal flexibility and capacity. Ostensibly, this uniformity is to reduce purchase, training, and maintenance costs, but fails to take into account risks and operational needs. Such attitudes are clearly nonsensical when applied to almost any other area of technology, so it is perplexing they are still rampant in IT.   

For instance, imagine buying a single model of commercial speedboat and assuming it will be adequate for bass fishing, auto ferries, arctic icebreakers, Coast Guard rescues, oil tankers, and deep water naval interdiction -- so long as we add on a few aftermarket items and enable a few options. Fundamentally, we understand that this is untenable and that we need to architect a vessel from the keel upwards to tailor it for specific needs, and to harden it against specific dangers. Why cannot we see the same is true for computing? Why do we not understand that the commercial platform used at home to store Aunt Bee's pie recipes is NOT equally suitable for weapons control, health care records management, real-time utility management, storage of financial transactions, and more? Trying to support everything in one system results in huge, unwieldy software on incredibly complex hardware chips, all requiring dozens of external packages to attempt to shore up the inherent problems introduced by the complexity. Meanwhile, we require more complex hardware to support all the software, and this drives complexity, cost and power issues.

The situation is unlikely to improve until we, as a society, start valuing good security and quality over the lifetime of our IT products. We need to design systems to enforce behavior within each specific configuration, not continually tinker with general systems to stop each new threat. Firewalls, IDS, antivirus, DLP and even virtual machine "must-have" products are used because the underlying systems aren't trustworthy--as we keep discovering with increasing pain. A better approach would be to determine exactly what we want supported in each environment, build systems to those more minimal specifications only, and then ensure they are not used for anything beyond those limitations. By having a defined, crafted set of applications we want to run, it will be easier to deny execution to anything we don't want; To use some current terminology, that's "whitelisting" as opposed to "blacklisting." This approach to design is also craftsmanship--using the right tools for each task at hand, as opposed to treating all problems the same because all we have is a single tool, no matter how good that tool may be. After all, you may have the finest quality multitool money can buy, with dozens of blades and screwdrivers and pliers. But you would never dream of building a house (or a government agency) using that multitool. Sure, it does a lot of things passably, but it is far from ideal for expertly doing most complex tasks.

Managers will make the argument that using a single, standard component means it can be produced, acquired and operated more cheaply than if there are many different versions. That is often correct insofar as direct costs are concerned. However, it fails to include secondary costs such as reducing the costs of total failure and exposure, and reducing the cost of "bridge" and "add-on" components to make items suitable. Smaller and more directed systems need to be patched and upgraded far less often than large, all-inclusive systems because they have less to go wrong and don't change as often. There is also a defensive benefit to the resulting diversity: attackers need to work harder to penetrate a given system because they don't know what is running. Taken to an extreme, having a single solution also reduces or eliminates real innovation as there is no incentive for radical new approaches; with a single platform, the only viable approach is to make small, incremental changes built to the common format. This introduces a hidden burden on progress that is well understood in historical terms -- radical new improvements seldom result from staying with the masses in the mainstream.

Therein lies the challenge, for researchers and policy-makers. The current cyber security landscape is a major battlefield. We are under constant attack from criminals, vandals, and professional agents of governments. There is such an urgent, large-scale need to simply bring current systems up to some bare minimum that it could soak up way more resources than we have to throw at the problems. The result is that there is a huge sense of urgency to find ways to "fix" the current infrastructure. Not only is this where the bulk of the resources is going, but this flow of resources and attention also fixes the focus of our research establishment on these issues, But when this happens, there is great pressure to direct research towards the current environment, and towards projects with tangible results. Program managers are encouraged to go this way because they want to show they are good stewards of the public trust by helping solve major problems. CIOs and CTOs are less willing to try outlandish ideas, and cringe at even the notion of replacing their current infrastructure, broken as it may be. So, researchers go where the money is -- tried and true, incremental, "safe" research.

We have crippled our research community as a result. There are too few resources devoted to far-ranging ideas that may not have immediate results. Even if the program managers encourage vision, review panels are quick to quash it. The recent history of DARPA is one that has shifted towards immediate results from industry and away from vision, at least in computing. NSF, DOE, NIST and other agencies have also shortened their horizons, despite claims to the contrary. Recommendations for action (including the recent CSIS Commission report to the President) continue this by posing the problem as how to secure the current infrastructure rather than asking how we can build and maintain a trustable infrastructure to replace what is currently there.

Some of us see how knowledge of the past combined with future research can help us have more secure systems. The challenge continues to be convincing enough people that "cheap" is not the same as "best," and that we can afford to do better. Let's see some real innovation in building and deploying new systems, languages, and even networks. After all, we no longer need to fit in 32K of memory on a $21 million computer. Let's stop optimizing the wrong things, and start focusing on discovering and building the right solutions to problems rather than continuing to try to answer the same tired (and wrong) questions. We need a major sustained effort in research into new operating systems and architectures, new software engineering methods, new programming languages and systems, and more, some with a (nearly) clean-slate starting point. Small failures should be encouraged, because they indicate people are trying risky ideas. Then we need a sustained effort to transition good ideas into practice.

I'll conclude with s quote that many people attribute to Albert Einstein, but I have seen multiple citations to its use by John Dryden in the 1600s in his play "The Spanish Friar":

"Insanity: doing the same thing over and over again expecting different results."

What we have been doing in cyber security has been insane. It is past time to do something different.
[Added 12/17: I was reminded that I made a post last year that touches on some of the same themes; it is here.]

[This is dervied from a posting of mine to Dave Farber's Interesting People list.]

There is an article in the October Businessweek that describes the problem of counterfeit electronic components being purchased and used in critical Defense-related products.

This is not a new threat. But first let's reflect on the past.

Historically, the military set a number of standards (MIL-SPEC) to ensure that materials they obtained were of an appropriate level of quality, as well as interoperable with other items. The standards helped ensure a consistency for everything from food to boots to tanks to software, as well as ensuring performance standards (quality).

The standards process was not without problems, however. Among issues often mentioned were:

• Standards were sometimes not revised often enough to reflect changes in technology. The result was that the military often had to acquire and use items that were generations behind the commercial marketplace (esp. in software/computers);
• Knowing and complying with so many standards often caused companies considerable extra time and effort in supplying items, thus raising their cost well above comparable commercial equivalents;
• Incompatible standards across military agencies and services, especially when compared with commercial items used by civilian agencies, led to waste and increased cost, and lack of flexibility in implementation;
• Imposition of rigid standards cut down on innovation and rapid development/acquisition/deployment cycles;
• The rigidity and complexity of the standards effectively shut out new vendors, especially small vendors because they could not match the standards-compliance efforts of large, entrenched defense vendors.

Thus, in June of 1994, William Perry, the then Secretary of Defense, issued a set of orders that effectively provide a pathway to move away from the standards and adopt commercial standards and performance goals in their place. (cf. the Wikipedia article on MIL-SPEC). One of the rationales expressed then, especially as regarded computing software and hardware, was that the competition of the marketplace would lead to better quality products. (Ironically, the lack of vendor-neutral standards then led to a situation where we have large monocultures of software/hardware platforms throughout government, and the resultant lack of meaningful competition has almost certainly not served the goals of better quality and security.)

In some cases, the elimination of standards has indeed helped keep down costs and improve innovation. I have been told, anecdotally, that stealth technology might not have been fielded had those aircraft been forced within the old MIL-SPEC regime.

As a matter of cost and speed many MIL-SPEC standards seem to have been abandoned to choose COTS whenever possible without proper risk analysis. Only recently have policy-makers begun to realize some of the far-reaching problems that have resulted from the rush to abandon those standards.

As the Businessweek article details, counterfeit items and items with falsified (or poorly conducted) quality control have been finding their way into critical systems, including avionics and weapons control. The current nature of development means that many of those systems are assembled from components and subsystems supplied by other contractors, so a fully-reputable supplier may end up supplying a faulty system because of a component supplied by a vendor with which they have no direct relationship. One notable example of this was the "Cisco Raider" effort from a couple of years ago where counterfeit Cisco router boards were being sold in the US.

As noted in several press articles (such as the ones linked in, above) there is considerable price motive to supply less capable, "grey market" goods in large bids. The middlemen either do not know or care where the parts come from or where they are being used -- the simply know they are making money. The problem is certainly not limited to Defense-related parts, of course. Fake "Rolex" watches that don't keep time, fake designer shoes that fall apart in the rain, and fake drugs that either do nothing or actually cause harm are also part of the "gray market." Adulteration of items or use of prohibited materials is yet another aspect of the problem: think "lead paint" and "melamine" for examples. Of course, this isn't a US-only problem; people around the world are victimized by gray-market, adulterated and counterfeit goods.

These incidents actually illustrate some of the unanticipated future effects of abandoning strong standards. One of the principal values of MIL-SPEC standards was that it established a strict chain of accountability for products. I suspect that little thought has been given by policy-makers to the fact that there is considerable flow of items across borders from countries where manufacturing expertise and enforcement of both IP laws and consumer-protection statutes may not be very stringent. Buying goods from countries where IP violations are rampant (If there is little fear over copying DVDs, then there is little fear over stamping locally-produced items as "Cisco"), and where bribes are commonplace, is not a good strategy for uniform quality.

Of course, there are even more problems than simply quality. Not every country and group has the same political and social goals as we do in the US (or any other country -- this is a general argument). As such, if they are in a position to produce and provide items that may be integrated into our defense systems or critical infrastructure, it may be in their interests to produce faulty goods -- or carefully doctored goods. Software with hidden 'features" or control components with hidden states could result in catastrophe. That isn't fear-mongering -- we know of cases where this was done, such as to the Soviets in the 1980s. Even if the host country isn't subtly altering the components, it may not have the resources to protect the items being produced from alteration by third parties. After all, if the labor is cheaper in country X, then it will also be cheaper to bribe the technicians and workers to make changes to what they are producing.

The solution is to go back to setting high standards, require authentication of supply chain, and better evaluation of random samples. Unfortunately, this is expensive, and we're not in a state nationally where extra expense (except to line the pockets of Big Oil and Banking) is well tolerated by government. Furthermore, this alters the model where many small vendors acting as middlemen are able to get a "piece of the action." Their complaints to elected representatives who may not understand the technical complexities adds even further pressure against change.

In some cases, the risk posed in acquisition of items may warrant subsidizing the re-establishment of some manufacturing domestically (e.g., chip fabs). This doesn't need to be across the board, but it does required judicious risk-analysis to determine where critical points are -- or will be in the future. We must realize that the rapid changes in technology may introduce new patterns of production and acquisition that we should plan for now. For instance, once elements of nanotechnology become security-critical, we need to ensure that we have sufficient sources of controlled, quality production and testing.

I'm not going to hold my breath over change, however. Some of us have been complaining about issues such as this for decades. The usual response is that we are making a big deal out of "rare events" or are displaying xenophobia. The sheer expense frightens off many from even giving it more than a cursory thought. I know I have been dismissed as an "over-imaginative academic" more times than I can count when I point out the weaknesses.

One of the factors that allegedly led to the decline of the Roman empire was the use of lead in pipes, and lead salts to make cheap wine more palatable for the masses. The Romans knew there was a health problem associated with lead, but the vendors saw more profit from using it.

Once we have sufficiently poisoned our own infrastructure to save money and make the masses happier, how long do we last?

[If you are interested in being notified of new entries by spaf on cyber and national security policy issues, you can either subscribe to the RSS feed for this site, or subscribe to the notification list.]

If you are in the United States, it has been nigh-on impossible to watch TV, read a newspaper, follow a blog, or (in some states) get your paper mail without something about the upcoming election being present. Some of this has been educational, but a huge amount of it has been negative, vague, and often misleading. That's U.S. politics, unfortunately -- the majority of voters don't really bother to educate themselves about the issues and the media does an uneven job of reporting the truth. For many voters, it comes down to only one or two issues they care passionately about, and they vote for a candidate (or against one) on those simple issues. For instance, there are many voters who will base their votes solely on a candidate's perceived position on gun control, access to legal abortions, tax policy, or other single issues without thinking about all the position issues. (And, as I note below, most of these single issues aren't really under the control of the President no matter who is elected.)

Of course, the US political system tends to reinforce this binary choice procedure, as we have long had only two really major parties. Parliamentary systems seem to encourage more parties, although even then there appears to be only two major ones, often oriented around the same approximate social/political poles: a conservative party, and a liberal (labor) party.

So, in the U.S. we have candidates from both major parties (and many minor ones) campaigning -- explaining their positions, offering their plans for when they are in office, and trying to instill voter confidence and trust. (And too often, offering innuendo, misquotes and outright untruths about their opponents.)

What none of them say, and the media doesn't either, is that very few of the promises can really be certain of being kept. And in large part, that is also a nature of government.

The President has a limited set of powers under the Constitution. He (or she) is responsible for the execution of the laws of the United States. The President is the Commander-in-Chief of all the armed forces and is responsible for commanding them in defense of the country and upholding the law (including treaties). The President is the chief executive agent of all the various Cabinet agencies, and of a number of offices and commissions. The President appoints a large number of officials (including judges and ambassadors), but doesn't have the power to remove many of them.

Most importantly, the President does not make new laws. Laws are passed by Congress, usually with the assent of the President, although a 2/3 majority of both houses of Congress may pass laws to which the President objects. The President is then responsible for ensuring that those laws are carried out, with recourse to the Courts if there are questions. If the President fails to enforce the laws, Congress may take some punitive actions, or even impeach the President...if they have the political will.

So, back to the candidates. If you listen to their speeches, they offer to change tax law, spend more on energy issues, change health care, and a number of other important domestic issues. What they don't point out, however, is that they will have no authority as President to do most of those things! Instead, Congress will need to pass authorizing legislation that is signed by the President. The President can certainly propose that Congress enact those changes, but Congress needs to craft and pass legislation that enables the President to act, and that allocate necessary funds, and that also create/remove administrative structures that may be involved. This legislation can include whatever other items that Congress adds in to the bill, including items that may be completely unrelated to the main topic. The President then must decide whether to sign the bill and act to implement its provisions.

So, the most a new President can do is to propose legislation to embody his/her campaign promises, and to work for its passage. What usually happens is that the size of the win in the election serves as a political measure of how much the population is aligned with the new President's positions, and this can help get a particular agenda passed...or not. Of critical importance is also the issue of whether one or both houses of Congress are controlled by the same party as the new President, and by what margin.

Thus, there should probably be more attention paid to the candidates running for Congress and their particular positions on important issues. In many venues, however, the majority of the attention is focused on the Presidential contest. Some other states are also dealing with contentious state initiatives, tight governor races, and other local issues that help further obscure the Congressional races.

Now, how does this apply to cybersecurity, the ostensible topic of this blog? Or education? Or privacy? Or other topics we focus on here?

Well, as I will address in my next posting, the two main Presidential candidates have made some comments on cyber security, but I have not been able to find any coverage of any current candidate for Congress who has mentioned it. It is basically invisible. So is privacy. Education has gotten a little mention, but not much. And given the more overt, pressing issues of the economy, the deficit, health care, energy dependence, and war in the Middle East, it seems unlikely that any Congressional candidate has bothered to think much about these cyber issues, or that they have received much further thought from the Presidential candidates. (Too bad cyber security can't be part of the mud slinging -- it might raise its profile!)

Of course, with the economy in such sad shape, and some of the other severe problems being faced by the U.S., one might ask whether cyber should be a priority for the new President. I would answer yes, because the problems are already here and severe (although not as obvious as some of the other problems), and it will take years of major effort simply to keep even with the current sad status. The problems in cyber cannot be fixed in a crash effort devoted at any future time, and until they are addressed they will be a drain on the economy (in 2006, the FBI estimated the loss to computer crime in the US to be $67 billion -- almost 10% of the recent economic bailout), and a threat to national security. Thus, deferring action on these issues will only make the situation worse; we need to initiate a sustained, significant program to make some important changes.

There are some things that the new President can do, especially as they relate to the military, law enforcement, and some other agencies in the Executive Branch. This is potentially cause for some glimmer of hope. I intend to blog some on that too, with a list of things that should be considered in the new administration.

Over the last month or so, many people who read my first post on Senator Obama's "security summit" at Purdue have asked me about followup, I've been asked "Did you ever hear back from the Senator?", "Has the McCain campaign contacted you?", and "What do you think about the candidates?" I've also been asked by a couple of my colleagues (really!) "Why would they bother to contact you?"

So, let me respond to these, with the last one first.

Why would someone talk with you about policy?

So, I haven't been elected or served in a cabinet-level position in DC. I haven't won a Nobel prize (there isn't one in IT), I'm not in the National Academies (and unlikely to be -- few non-crypto security people are), and I don't have a faculty appointment in a policy program (Purdue doesn't have one). I also don't write a lot of policy papers -- or any other papers, anymore: I have a persistent RSI problem that has limited my written output for years. However, those aren't the only indicators that someone has something of value to say.

As I've noted in an earlier post, I've had some involvement in cyber security policy issues at the Federal level. There's more than my involvement with the origins of the SfS and Cyber Trust, certainly. I've been in an advising role (technology and policy) for nearly 20 years with a wide range of agencies, including the FBI, Air Force, GAO, NSA, NSF, DOE, OSTP, ODNI and more. I've served on the PITAC. I've testified before Congressional committees a half-dozen times, and met with staff (officially and unofficially) of the Senate and House many times more than that. Most people seem to think I have some good insight into Federal policy in cyber, but additionally, in more general issues of science and technology, and in defense and intelligence.

From another angle, I've also been deeply involved in policy. I served on the CRA Board of Directors for 9 years, and have been involved with its government affairs committee for a decade. I've been chair or co-chair of the ACM's US Public Policy committee for a dozen years. From these vantage points I have gained additional insights into technology policy and challenges in a broad array of issues related to cyber, education, and technology.

And I continue to read a lot about these topics and more, including material in a number of the other sciences. And I've been involved in the practice and study of cyber security for over 30 years.

I can, without stretching things, say that I probably know more about policy in these areas than about 99.995% of the US population, with some people claiming that I'm in the top 10 or so with respect to broad issues of cyber security policy. That may be why I keep being asked to serve in advisory positions. A lot of people tend to ask me things, and seem to value the advice.

One would hope that at least some of the candidates would be interested in such advice, even if not all of my colleagues (or my family :-) are interested in what I have to say.

Have any of the other candidates contacted you?

Simply put -- no. I have gotten a lot of mailings from the Republican (and Democratic) campaigns asking me to donate money, but that's it.

I'm registered as an independent, so that may or may not have played a role. For instance, I can't volunteer to serve as a poll worker in Indiana because I'm not registered in one of the two main parties! I don't show up in most of the databases (and that may be a blessing of sorts).

To digress a moment.... I don't believe either party has a lock on the best ideas -- or the worst. I'm not one of those people who votes a straight-ticket no matter what happens. I have friends who would vote for anyone so long as the candidate got the endorsement of "their" party. It reminds me of the drunken football fans with their shirts off in -20F weather cheering insanely for "their" team and willing to fight with a stranger who is wearing the wrong color. Sad. Having read the Constitution and taken the oath to defend it, I don't recall any mention of political parties or red vs. blue....

That said, I would be happy to talk with any serious candidate (or elected official) about the issues around cyber, security, education, and the IT industry. They are important, and impact the future of our country...and of much of the world.

So, has anyone with the Obama campaign contacted you since his appearance at Purdue?

Well, the answer to this is "yes and no."

I was told, twice, by a campaign worker that "Someone will call you -- we definitely want more advice." I never got that phone call. No message or explanation why. Nothing.

A few weeks after the second call I did get a strange email message. It was from someone associated with the campaign, welcoming me to some mailing list (that I had not asked to join) and including several Microsoft Word format documents. As my correspondents know, I view sending email with Word documents to be a bad thing. I also view being added to mailing lists without my permission to be a hostile act. I responded to the maintainer of the list and his reply was (paraphrased) "I don't know why you were added. Someone must have had a reason. I'll check and get back to you." Well, I have received no more email from the list, and I never got any followup from that person.

So, in summary, I never got any follow-up from the campaign. I don't think it is an issue with the Senator (who wouldn't have been the one to contact me anyhow) but a decision by his staff.

So, depending your level of cynicism, the mentions of my name, of CERIAS, and of follow-up was either (a) a blown opportunity caused by an oversight, or (b) a cynical political ploy to curry local favor.

(My daughter suggested that they are waiting until after the election to appoint me to a lofty position in government. Uh, yeah. That probably explains why I haven't gotten that MacArthur "genius grant" yet and why Adriana Lima hasn't called asking me to run away with her -- the timing just isn't right yet. :-)

What are your opinions on the Presidential candidates?

I'm not allowed to be partisan in official Purdue outlets. So, in some further posts here over the next week or two I will provide some analysis of both major candidates (NB. Yes, I know there are over 300 candidates for President on the ballots across the country. However, I don't think there is much chance of Baldwin, Barr, McKinney, Nader, Paul or the rest getting into office. So, I'll limit my comments to the two main candidates.

If you really want to know who I'm probably voting for, you can see my Facebook page or send me email.

---

[Update 10/16: After this was published I sent a link to this entry to several people associated with the Obama campaign. Only one responded, and it was clear from his email that there had been a mixup in getting back to me -- but no interest in rectifying it.]

In my secure programming class I have denounced the overloading of return values as a bad practice, and recently I discovered a new, concrete example of a bug (possibly a vulnerability) that results from this widespread practice. A search in the NVD and some googling didn't reveal any mention of similar issues anywhere, but I probably just have missed them (I have trouble imagining that nobody else pointed that out before). In any case it may be worth repeating with this example. The practice is to return a negative value to indicate an error, whereas a positive value has meaning, e.g., a length. Imagine a function looking like this:

int bad_idea(char *buf, unsigned int size) {
    int length;

    if (<some_error_condition>) {
        length = -ERROR_CODE;
    } else {
        length = size;  // substitute any operations that could overflow the signed int
    }
    return length;
}

This function could return random error values. Under the right conditions this could result in at least a DoS (imagine that this is a security-related function, e.g., for authentication). I suggest using separate channels to return error codes and meaningful values. Doing this lowers complexity in the assignment and meaning of that return value by removing the multiplexing. As a result:

• There is an increased complexity in the function prototype, but the decreased ambiguity inside the function is beneficial. When the rest of the code uses unsigned integers, the likelihood of a signed/unsigned integer conversion mistake or an overflow is high. In the above example, the function is also defective because it is unable to process correctly some allowed inputs (because the input is unsigned and the output needs to be able to return the same range), so in reality there is no choice but to decouple the error codes from the semantic results (length). This discrepancy is easier to catch when the ambiguity of the code is decreased.
• It does away with the bad practice of keeping the same variable for two uses: assigning error codes and negative values to a "length" is jarring;
• It disambiguates the purpose and meaning of checking the "returned" values (I'm including the ones passed by reference, loosely using the word "returned"). Is it to check for an error or is it a semantic check that's part of the business logic?

Integer overflows are a well-known problem; however in this case they are more a symptom of conflicting requirements. The incompatible constraints of having to return a negative integer for errors and an unsigned integer otherwise are really to blame; the type mismatch ("overflow") is inevitable given those. My point is that the likelihood of developers getting confused and having bugs in their code, for example not realizing that they have incompatible constraints, is higher when the return value has a dual purpose. Edit (10/13): It just occurred to me that vsnprintf and snprintf are in trouble, in both BSD and Linux. They return an int, but take an unsigned integer (size_t) as a size input, AND are supposed to return either the number of printed characters or a negative number if there's an error. Even if the size can be represented as a signed integer, they return the number of characters that *would* have been printed if the size was infinite, so specifying a small size isn't a fix. So it *might* be possible to make these functions seemingly return errors when none happened. Note(10/14): I've been checking the actual code for snprintf. In FreeBSD, it checks both the passed size and the return value against INT_MAX and appears safe.