Hacker Newsnew | past | comments | ask | show | jobs | submit | more lambda's commentslogin

> Whether or not this is the case, I don't know, but it kinda feels like running your app on a shared-hosting account...

Remember that you are running your app on a shared hosting account with Heroku.

Now, it has somewhat more isolation than a typical shared hosting account, but less isolation than using physically separate hardware.

Heroku is built on top of AWS, so your machine is running on the same physical hardware, though a different virtual machine, as other AWS customers. Furthermore, Heroku uses LXC to isolate its dynos; so you are running on the same VM as another Heroku customers, albeit separated by a container barrier. And finally, if you're using Postgres, then you're running on a shared database service, which is pretty much exactly like what you'd get with shared hosting.


When I mined my first block of bitcoins a couple years ago (back when I could use the GPU on my aging laptop and still have a chance to mine a block) they were worth about 5¢; I sold a bunch when they were around $5 because I thought that was ridiculously high, and the price was sure to collapse. Silly me; had I not sold them, I could have a few thousand dollars from a couple of days worth of GPU time.


They require that you have the ability to change the keys; remove the Microsoft ones, load your own keys, etc.

The problem is that they don't specify any kind of common interface for doing so, or standard for key formats, or anything of the sort. So how it works can vary greatly between manufacturers, or may be buggy, or the like. So several prominent Linux distros have decided to ship bootloaders signed with Microsoft's keys, so that people can just install the distro without having to go through some hardware specific key loading process that may brick their machine.


But you can easily force regular DNS to give you high amplification. Just query a domain which gives you a larger response, such as multiple A records, MXes, SPF records, and so on. For example:

    $ dig google.com. ANY | grep SIZE
    ;; MSG SIZE  rcvd: 546
If you want more amplification than that gives, just host one yourself. The recursive resolver will hit your DNS server once, then send out replies based on the cache.

There seems to be lots of fearmongering about DNSSEC amplification, but you can get just the same amount of amplification out of regular DNS, so it seems that fixing DNS amplification in other ways would be more effective than trying to avoid adopting DNSSEC.


> There seems to be lots of fearmongering about DNSSEC amplification

There is also a 300+ Gbps DDoS attack making use of it right now. This was foreseen as a huge amplification vector in a stateless protocol during the design phase, but was ignored. Now we get to reap the benefits of that decision.

Normal DNS responses don't often grow to the size of google.com/IN/ANY. You have a very limited number of authoritative sources to use (and hosting yourself creates a bottleneck as well as a path back to you). With widespread DNSSEC adoption every zone becomes a good amplification source, which nullifies the current best practices for mitigation (rate limiting responses per zone/source).

If DNSSEC adoption becomes the norm, open recursive resolvers no longer become the problem and direct to authoritative becomes a viable attack vector.


  | There is also a 300+ Gbps DDoS attack
  | making use of it right now
I have yet to see anyone state authoritatively that DNSSEC is being used in this attack. Could you provide a reference for this?

If this attack right now is able to reach 30x amplification without DNSSEC, then what's the point of of decrying DNSSEC amplification as a huge issue?

Other discussion: https://news.ycombinator.com/item?id=5451299


DNSSEC is the amplification in "DNS amplification attack." I personally run a (heavily rate limited) open resolver as a honey pot to observe these attacks in progress.

You can read CloudFlare's own explanation of how these attacks work http://blog.cloudflare.com/deep-inside-a-dns-amplification-d...


Unless I'm misunderstanding, the 'amplification' in 'DNS amplification attack' doesn't necessarily refer to DNSSEC. The idea is that you use x amount of bandwidth to send y amount of bandwidth at the target where y = kx, for some value of k that is significant enough to make it more worthwhile than just sending the traffic directly.

E.g. make a UDP DNS request to an open resolver with the source IP forged to be your target, then the response is sent to your target (rather than to the real source of the request).

My understanding is that the problem people have with DNSSEC in this regard is that the data returned in those responses increases by a lot (allowing for a 30x increase?). But if attackers are able to accomplish this without DNSSEC, then what's the point of talking about how horrible DNSSEC will make things in this regard?


So could some right thinking person scan the internet for open DNS resolvers and perform DDoS attacks against them using other open resolvers?


Yeah! Right Thinkers! DDoS all those awful Wrong Thinkers! Knock their bastard servers right off the internet! That's the solution!


It's actually quite an elegant solution to getting people o configure their servers correctly. Instead of their servers being a hazard to the wider internet, they become a hazard to each other.


Do you simply not care about the collateral damage, or do you feel the ends justify the means?


I used to be staunchly against deletionism, but having had more experience with exactly the above kind of page, where someone who cares a lot about some organization or topic that really isn't at all notable can go to an awful lot of trouble to create stuff that just junks up Wikipedia with biased, poorly researched and poorly cited pages, I am a lot more sympathetic to deletionism these days.

There are real costs to maintaining pages; editors need to revert vandalism, fix it to match the style guidelines, look for citations, resolve disputes, keep it neutral, ensure that it reflects a broad consensus about the actual truth of the matter, and so on. For small, insignificant topics, an article can be more trouble than it's worth.


You could always stub and protect such a page instead of deleting it.

I don't care about orgs but when a band or programming language bages get deleted i'm not happy.


destruction is selection, and selection is destruction


I'm not sure why you think the Wikipedia policy is wrong.

If there is no substantial third party coverage of a topic, why should Wikipedia cover it?

A lot of people seem to believe that since bits are cheap, we might as well keep around any page no matter how non-notable it is. But there is a cost to including articles beyond the storage cost. Articles need to be patrolled for vandalism, and fact checked based on material published elsewhere. If you let people just post anything they want, then people will start using it for spam, fluff pieces, random bullshit, and so on. Enough of that, and it will substantially tarnish Wikipedia's reputation as a useful and unbiased source of information.

If a person or organization has never received press coverage, academic research, or other substantial information published on them outside of Wikipedia, why should they have a Wikipedia page?

Should every Joe Blow who happened to be captain of the swim team in high school and who's now the local Rotary Club president have a Wikipedia page? Should every mom and pop pizza joint? Who gets to have the Wikipedia page for "Village Pizza"?

The only reason for these small, not otherwise notable organizations to have a Wikipedia page is to increase their perceived notability; having a Wikipedia page indicates a certain amount of fame or notoriety. So in that case, allowing not otherwise notable organizations to have Wikipedia page would basically be a way for them to get free advertising. Especially in a case like this, when the whole page was edited by people who are, apparently, closely connected to the organization in question, and who would fight any negative information placed on the page.


If there is no substantial third party coverage of a topic, why should Wikipedia cover it?

Because "substantial third-party coverage" is a flawed test for the importance of a subject.

Press coverage tends toward the sensational, visual, beautiful, controversial, current, language-specific, and easily explained. If a topic doesn't meet those criteria, it probably won't be covered by the press -- unless the topic in question has some well-connected PR firm or publicist pushing for it.

Another issue with press coverage is "established" media is contracting. There aren't as many editors assigning stories, or reporters crafting unbiased profiles of companies, organizations, topics and people. This has an significant, indirect impact on the types of topics that can be sourced according to Wikipedia's official rules.

Meanwhile, the number of self-published, user-generated sources of expert opinion is exploding. I'm talking about blogs and places like Hacker News and Reddit. Someone with deep experience and understanding of a topic can share information, yet this information can't be used, according to Wikipedia's policies.


> Because "substantial third-party coverage" is a flawed test for the importance of a subject.

Its not, really, for something that by-mission is a tertiary source. There's probably a different set of rules that ought to be in place for a publicly-contributed-to Wikijournal of Original Research (which, once it established itself as credible, could be a source cited in Wikipedia), but that's a different niche. Wikipedia does not intend to be all things to all people, it intends to be one thing and do it well.

> Press coverage tends toward the sensational, visual, beautiful, controversial, current, language-specific, and easily explained.

> Another issue with press coverage is "established" media is contracting.

"Third-party" and "press" aren't the same thing. Wikipedia, by policy, prefers academic and peer-reviewed publications.

News media publications are one of many other things that might be reliable sources in particular areas, but are not especially preferred.

> Meanwhile, the number of self-published, user-generated sources of expert opinion is exploding. I'm talking about blogs and places like Hacker News and Reddit. Someone with deep experience and understanding of a topic can share information, yet this information can't be used, according to Wikipedia's policies.

It actually can, per Wikipedia policy, if the author of the self-published source also has other work in the relevant field published in reliable, third-party publications.


> Because "substantial third-party coverage" is a flawed test for the importance of a subject.

So what? Importance of a subject is not what WP is trying to filter. Wikipedia wants to contain all the important knowledge, yes, but that's impossible to measure or agree on, so there is a set of criteria that act as a reasonable proxy, and you need to accept that the Wikipedia we've got is the one that has those filters.

So it goes something like this: content in wikipedia should be "NPOV", it should not be original research (i.e. WP aims to be a tertiary source), and content should be "verifiable". These terms in turn need to be refined, defined, given guidelines, etc. (For example, nothing is truly NPOV, so we need a way to understand how to approximate it, and for example, not all verification is equal (and nothing is perfectly verifiable), so we need guidelines again.)

And, as sibling comments have pointed out, there is NOT a preference for "press coverage" and "established media". It's one permitted alternative.

> "expert opinion ... like Hacker News and Reddit"

I assume that you were joking, or making some nuanced point that I missed, but just in case you weren't...

Listen, they let ME write comments on Hacker News and Reddit, you know. They also let experts write at such places, it's true, but how will you know the difference? Yes, there's lots of expert opinion in the world, including e.g. some on HN, but so what? Citations aren't any good if they don't bring some kind of authority, and "some guy said so on HN" is not very convincing.

There are certainly circumstances in which a comment on HN can qualify for WP:RELIABLESOURCES .

There's lots of good reading at http://en.wikipedia.org/wiki/Wikipedia:About , but if you're trying to "get something done" today, it's too much to read at one sitting. You've gotta just accept that, but then again you shouldn't be trying to "get something done" on wikipedia in a given day.


I agree press coverage is suboptimal. But really, press coverage is one of the lower tiers of "reliable source" from Wikipedia's perspective, and cautiously included for mostly pragmatic reasons. Ideally, articles shouldn't be based on press coverage: if you're writing an article about the history of IBM, or about World War I, you should base it on one of the many books or journal articles on the subject, not on digging up some old New York Times article. That's precisely because it's actually not always the case that whatever the NYT happened to write at the time is either accurate or a complete picture. It's better to cite a proper historical study that's done a critical analysis of the original sources instead.

But for very recent things where nobody has written anything better, newspapers are treated as an acceptable source. Ideally, they will be replaced with something else, at some later date: the Syrian War article is currently mostly sourced from newspapers, but in 30 years it really should be written with citations to major books and articles on the subject, not cobbled together from newspaper articles. There are a few exceptions; for example, longer-form magazine biopics, and newspaper obituaries, sometimes provide good longer-term sources for middle-level-fame figures who aren't well-known enough for anyone to have written books or journal articles about them.


> There's legitimate substance here, but occasionally, the system actually works exactly the way it's supposed to -- the judicial branch both asserting its independence from the executive and striking down an overreaching action by the latter.

Note that this isn't over yet. The judge put a stay on the ruling to allow the government to appeal. Note that national security letters have been around since 1978; and have been greatly expanded in scope in 2001. So depending on how you count, that's 12 years or 35 years of lost liberty. I'm not sure I would count that as a system working the way it's supposed to. How many years of living under gag orders do we consider acceptable?


Yep. Donate to the EFF regularly. With all luck, you won't need their services; but in the rare case that you do, it's much better to have them still be around and well funded.


Why is it that so many of these alternatives are hosted web apps? Isn't the lesson to learn from Google Reader going away that you might not want to trust hosted web apps for the long term? Why aren't there more client-side app alternatives?


Number of times google reader has a planned shut down, in which I still have plenty of time to grab my data: 1

Number of times something has happened to make an old desktop app stop working: numerous

for example (I haven't experienced all of these):

- switching from Windows -> Mac - having my hard drive fail and finding out my backups were broken - OSX upgrades breaking backwards compatibility (ie, old binary format) - having my laptop stolen - switching jobs, requiring me to give my laptop back - having multiple devices that I like to read news on (ie, phone, laptop, tablet)

I much prefer to have as little as possible tied to the hunk of metal my sweaty fingers type on. I haven't lost any data yet from using 'cloud services' and the GR replacement I pick will be well vetted.


I value my privacy far more than any feature or convenience a web app may offer. This is one of the main reasons why I stay away from web apps as much as possible, and use open-source client-side apps whenever I can. As a side-benefit, I also don't have to rely on the whims of the web app provider to keep their service running.

Unless you use web apps for absolutely everything, you're still going to have to make backups. And if you're making backups anyway, how hard is it to include a tiny little list of RSS feed URLs which can be used to pick up from where you left off with any RSS reader, should anything happen?

Anyway, if you screw up your backups, you've got far more serious problems than simply restoring an RSS feed list. Do a good job on backups, and the rest mostly falls in to place.


I don't use a client-side reader because one of my use cases is subscribing to a lot of feeds that have dozens of updates a day. Most of those feeds typically only show the last N items, where N is significantly less than the number of posts in a day.

If I close my laptop for the weekend, that means I lose most of those entries in my reader.


If you use OS X, I can recommend Vienna, which has no problem at all with pulling and keeping as many feeds and articles as you want, up to as many as your filing system can cope with. It's mature, having been around for a few years now, and open source to boot.


The problem isn't running out of disk space, the problem is that when my computer is not running, it's going to miss some RSS entries.


Hmm. I personally have needed to use Google Code Search several times since it shut down, but have never found your site. When I've looked for alternatives by searching Google (say, for "code search"), I've found Koders (now Ohloh code search), but not yours. It seems like you might need to do a bit more SEO.


Yes I certainly do. I have been working on it, but Koders and Ohloh are pretty established so its a long climb. I think the .de domain might be hurting a little bit too. Happy to listen to any actionable items that can address this though.


Check out the periodic table of SEO: http://searchengineland.com/seotable

Kalzmeus on SEO (click around his blog for more, guy's a genius): http://www.kalzumeus.com/2010/01/24/startup-seo/

Also some quick, simple advice:

1. Blog 2/3 times per week minimum. Target a single, relevant keyword per blog post (relevant to your domain). The post title, H1s, H2s etc, image tags should all contain your primary keyword. Blog posts don't have to be long, but they do have to be useful.

2. Build a G+, FB, Twitter account for your site if you haven't already. Blast out your new posts to these accounts using Hootsuite (free, but you can pay if you like).

3. Build an audience on the social networks - just search around, follow relevant people and share useful, educational material. Don't spam.

4. Personalised mass email from Google Docs and Gmail (kind of hesitant to recommend this in light of Google shuttering services!): http://www.labnol.org/internet/personalized-mail-merge-in-gm... you need to have "newsletter" signs ups (blast out a digest of your blog posts once per month), calls to actions etc on your site for this to work properly (i.e. optimise your site to capture names & emails). Great, low cost/free way to make sure your audience stays engaged and coming back to your site. You can pay for this kind of service too, e.g. Mailchimp (they've a free tier), Constant Contact, or the likes of Hubspot for full suite of marketing automation tools.

5. Best of Startup marketing/SEO here, that I've compiled over the years (absolute gold): https://docs.google.com/spreadsheet/ccc?key=0Ag_fyIIMSJ6DdGt...

6. Check out Sell More Software by patio11: http://www.hyperink.com/Sell-More-Software-Website-Conversio...

Many engineers don't fully appreciate marketing as patio11 would be quick to point out, but there's a start.


Being successful at SEO is almost the same as starting your own company. Any one of these is a full-time proposition.

I would seek out professional SEO company's and try to find the best deal. No way anybody could do all this stuff by themselves. In the end, it will be a worthwhile investment. From my own experience it's very time consuming to get it right.


Agreed. I do most of the above as an employee in a startup company, and we use Hubspot to have most of our marketing efforts "under one roof".

However if you're a bootstrapped company, you may not have a choice but to put the time in yourself rather than paying for professional SEO services.

If you have a technical and non-technical co-founders, this list is a large chunk of what the non-technical guy should/will be doing.

Building a great product that people want is pointless without letting people know about it.


@patrickk I've been looking for a basics post about SEo without tech speak for ahem years. That doc is brilliant. Thank you! I might be able to start implementing now instead of having SEO on my to-do list yet again.


Glad you found it helpful!


This is excellent, thanks so much!


Not a huge thing, but your HTML isn't valid - for example, there's Google analytics stuff after the closing </html> tag.

That kind of thing is never going to help

http://validator.w3.org/check?uri=http%3A%2F%2Fsearchco.de%2...


I didn't know that had any bearing on it but thinking about it seems like it would be a ranking factor. I'll fix that up soon and see if anything changes. Cheers.


That doesn't make any difference at all to SEO (and in some cases can improve performance - or at least having it after the body tag does).


Grepcode also turns up frequently in results.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: