The solution is NOT simple. I suppose it always seems simple to impose rules on a system you don't understand.
Thousands of systems would need to change at brokers, banks, clearing firms, trade processing firms etc including mainframes running Cobol software that was written in the 70's at quasi-governmental firms like the DTC. Documentation for these systems is mainly non-existent or incorrect.
What happens now is that at the end of the day registered brokers net out all trades done by them and send files which include trade details and net trades to the DTC. On T+1, you and T+2 you figure out what would fail and why. For example, let's say Zuckerberg sells some FB. Based on FB's corporate bylaws his shares change class when he sells them so they have fewer voting rights. So when he sells them the system at his brokerage barfs because the cusip does not match the publicly traded share cusip which then leads to a cascade of fails down the line because whoever he sold it to may have sold it to someone else.
You can't just automate rules like change the class of stock if Zuck sells it because the rules are different for every company and the rules change as the board adopts new bylaws and companies change their capital structure and the systems can't change that quickly because they have to be super, super reliable. If the DTC's clearing systems go down, the US economy has a BIG PROBLEM.
About 1% of trades result in a potential failure to deliver on T+1. The vast majority of these are resolved between T+1 and T+3 so that the trade clears. Some don't. It could be something like an investor who still holds paper certificates delivering them to the broker but the courier was late etc. Resolving these fails often involves the company's registrar in addition to financial systems, for example FB's registrar would have to change the class of stock if Zuck sold his stock.
Company registrar's aren't exactly paragons of customer service. Companies generally want to hire the cheapest firm because to them it's just an added arcane expense. The registrar provides service to a back office guy at a broker, and he back office guy really doesn't have a channel to complain to the company's CFO. So the registrar generally has no incentive to provide good service and the company has no reason to fire them and change them because as far as the company is concerned it's just a expense they want to minimize and changing registrars is a huge headache.
A failure to deliver does not mean the trade is somehow cancelled. It means that either the broker or the seller needs to find stock that is good to deliver. If you go out to T+10 and there isn't a VERY GOOD DOCUMENTED reason for the fail, like a tardy registrar, the buying broker and DTC will demand deliver and buy you in if necessary, ie buy the stock in the open market and deduct your account for the cost including a generous commission for themselves. So a fail is generally a very bad thing that a seller and his broker really, really want to avoid.
Pre-2008 the rules around short selling were that if a stock was widely available for borrow, you could short it without checking. Widely available means many different owners willing to lend the stock at reasonable prices.
In 2008 in a totally made up frenzy the, same SEC which failed to stop Madoff decided to impose a new rule stating that you needed affirmative determination of a borrow before you shorted a stock. The rule makes no sense because if there were to be an event that would cause a widely available stock to suddenly become not available, you would still have fails. Remember settlement is T+3, so you would check the accounts of the people willing to lend the stock and see that they still had the stock there without realizing that it would be gone in 3 days.
The SEC subpoenaed pretty much everyone because they decided that of course the market and particularly financial stocks were going down due to naked shorting as opposed to uh'm the economy and found essentially no evidence of manipulative short selling. They brought one case against a tiny broker who made a minor paperwork error essentially unrelated to affirmative borrow other than a truly tortured link link between the two made in the SEC's press release.
Short selling is a GOOD THING. Remember in any market, the people who have the biggest influence on prices are the ones who have amassed the most capital by consistently being right. When someone who has been consistently right decides to sell a stock short, s/he is providing a very valuable service to the economy by signaling that what she's shorting is overpriced. Remember shorting is very risky. The price of a stock has no upper limit but can only go as low as 0. So when you short you can make at most the current price of the stock but lose an infinite amount. Also remember that when you short, you also have to buy the stock back at some point. There are those pesky people you borrowed the stock from who want you to return it. So on average you are selling the stock when it's overpriced and buying it back when it's underpriced and the price is under pressure. The overall mechanism creates a very valuable self correction mechanism that is absolutely essential for a healthy market. If the SEC really cared about the market's health as opposed to PR, they would be working to make shorting easier because it prevents bubbles.
Remember that the total amount of short interest in a stock is a publicly available number. If the stock of a company you are an insider of is heavily shorted and you think it's cheap and you actually believe it as opposed to want to create a huge hullabaloo, you can go out and buy it, make noise about the fact that you are buying it. If the stock is genuinely cheap or even fairly priced, you will generally create a short squeeze where shorts are forced to buy back quickly as the stock shoots up in their face. OSTK insiders did not do that.
I don't actually know of a single instance where an insider or corporate inside has complained about a short seller shorting their stock and pushing the price down where there actually weren't genuine problems with the company. Remember it often takes 2-3 years for the whole story to come out and by then the media may not want to acknowledge that they breathlessly repeated every bit of tripe the CEO said about the evil, faceless short sellers. Usually a CEO complaining or a company filing a lawsuit against a short seller is a very high probability sign that the company is in genuine trouble because otherwise you would see insiders backing up the truck to buy the stock.
I find it truly amazing that the press and from the comments even the supposedly informed set at Hacker News would get worked up about this when it's about OSTK, Overstock.com. They sued people for shorting the stock and driving the price down for 60 dollars a share!! by making negative statements about the company that weren't true. With the benefit of hindsight it's clear that pretty much everything negative publicly said about OSTK was not just true but probably too optimistic, and the management has been lying which is felony under US law. Yet the SEC refuses to prosecute the management which is who is truly responsible for fleecing gullible retail investors.
It's truly astounding that OSTK is now choosing to go after GS based on snippets of emails etc which sound bad if you don't understand financial terminology and also don't have context but really aren't anything particularly bad, unless of course you let Tabibi self-aggrandize himself by claiming he understands stuff he has absolutely no clue about.