TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

High-Speed Trading Firm Deleted Some Code by Accident

75 点作者 fahimulhaq超过 9 年前

14 条评论

smoothgrammer超过 9 年前
In finance you&#x27;ll find low accountability for developers. Basically finance is where great developers go to get paid a lot to become horrible developers. Very few companies in finance are pushing the bar on quality code since they culturally cannot accept the cost of that. It doesn&#x27;t make sense until it is too late and by then their code base is too entrenched.<p>I&#x27;ve seen this at countless finance companies, and from interviewing people across the entire spectrum of finance. It is a rare corner of the industry where you&#x27;ll find quality engineering. Usually everyone in a finance company thinks they are amazing engineers, and are just plain wrong.<p>My advice: work smarter, not harder.
评论 #10320013 未加载
评论 #10317748 未加载
评论 #10318063 未加载
herbig超过 9 年前
The best part of this article is the feature graphic, that of a castle with the caption:<p>THIS IS A TOWER. PHOTOGRAPHER: DAMIEN MEYER&#x2F;AFP&#x2F;GETTY IMAGES<p>The article mentions &quot;Tower Research&quot;.<p>Is the image&#x2F;caption a joke, auto-generated by content, or just plain laziness?
评论 #10317053 未加载
评论 #10320479 未加载
x0054超过 9 年前
Every time I read stories about high speed trading, I have to wonder what is actually the point of high speed trading to the society. The point of a free and fast stock market is to provide monetary liquidity and stabilize the economy by shortening the feedback loop. But something tells me that our society and economy derives zero benefit from nanosecond resolution trading vs. trading with say 1 minute resolution. Meanwhile, the market is exposed to wild swings which might be smoothed over or even prevented if markets worked a little slower.
评论 #10317956 未加载
评论 #10317947 未加载
评论 #10317163 未加载
评论 #10319545 未加载
keyle超过 9 年前
Doesn&#x27;t matter how much money is involved or the quality of a team, someone somewhere will always do an oopsie. Maybe that particular &#x27;code&#x27; should have had a little comment such as &#x2F;&#x2F; used by ISO stuff do not delete<p>Commenting tricky &#x2F; non-obvious code is well recommended.
评论 #10316406 未加载
jwatte超过 9 年前
Acceptance testing fail! When changing code, and especially code that has consequences, doing so without tests is a bad idea. Doing so to protocol code without conformance tests is doubly bad.
评论 #10316850 未加载
tbrownaw超过 9 年前
...I&#x27;m missing something here. How do they not know how much they&#x27;ve asked to trade and not heard back on yet? Wouldn&#x27;t they <i>have</i> to track that, to not try to trade the same exact shares the same exact way a bunch of times? Is their local system maybe just more synchronous than I&#x27;m assuming would be possible?
评论 #10316569 未加载
Sander_Marechal超过 9 年前
Can someone explain <i>why</i> the SEC rules insist you <i>must</i> buy from the exchange with the lowest price first? It appears to me to be a rule just to keep the small exchanges alive.
评论 #10317874 未加载
评论 #10317924 未加载
fahimulhaq超过 9 年前
On a related note, Knight Capital lost $440 million due to a software bug a few years ago.<p><a href="http:&#x2F;&#x2F;www.bloomberg.com&#x2F;news&#x2F;articles&#x2F;2012-08-02&#x2F;knight-has-all-hands-on-deck-after-440-million-bug" rel="nofollow">http:&#x2F;&#x2F;www.bloomberg.com&#x2F;news&#x2F;articles&#x2F;2012-08-02&#x2F;knight-has...</a>
评论 #10324116 未加载
fahimulhaq超过 9 年前
Given hindsight, it&#x27;s always easier to ask for more tests or code comments. I guess the rule of thumb is that if the code is there, it most probably is there for a reason.<p>When something like this happens, it&#x27;s always a series of oversights.<p>1. Someone forgot to add comments or test cases while writing the code (or perhaps he wrote it but people are not running the test cases before commits).<p>2. Someone else thought that the code is dead and deleted it (and skipped the test cases if there were some covering that particular scenario).<p>3. Either the person deleting the code didn&#x27;t wait for the code review or code reviewers missed that as well. (If code reviews are non-existent, it&#x27;s a disaster waiting to happen).<p>So in the hindsight, write comments, write test cases, run test cases and do code reviews carefully.
pjc50超过 9 年前
Move fast. Break things. Pay millions of dollars in SEC fines.
osullivj超过 9 年前
A change like this making it through to production is indicative of a test failure, either automated or human. I have had the luxury of substantial human test teams on trading systems before 2008, and it was great. They wouldn&#x27;t let something like this through...
JimDash2145超过 9 年前
Who doesn&#x27;t notice a huge deletion in one changeset? This seems to be a convenient excuse. The SEC report.<p><a href="http:&#x2F;&#x2F;www.sec.gov&#x2F;litigation&#x2F;admin&#x2F;2015&#x2F;34-76029.pdf" rel="nofollow">http:&#x2F;&#x2F;www.sec.gov&#x2F;litigation&#x2F;admin&#x2F;2015&#x2F;34-76029.pdf</a>
Tinyyy超过 9 年前
I don&#x27;t know much about stocks trading, can someone ELI5? Thanks.
评论 #10316861 未加载
zeeshanm超过 9 年前
Another reason code needs to be idiomatic.