If I hadn't seen this article on Hacker News today, is there any combination of plausible keywords I could have typed into Google to find this particular article if I were looking for an in-depth review of microcontrollers?
I tried "best microcontrollers reviewed", "best low-cost microcontrollers", "survey of microcontrollers", and a bunch of others. Nothing that Google finds in the first 10-20 hits is as good this resource, other than Wikipedia, though very good, isn't a technical review.
I suppose the lessons are: there's room for lots of improvement in search engines, or SEO really works and you better promote yourself no matter how good you are, or word of mouth (Hacker News in this case) is still the way to find the best stuff.
This article is remarkable because there is typically very limited accessible blogs about topics like this. This long article requires in-depth knowledge of a large number of microcontrollers, and the time, desire, and ability to verify, collate, and summarize all that information. That's a rare combination.
An ordinary journalist could not write this article. It would take hundreds or thousands of hours to gain the domain knowledge required to understand what needs to be written up and described - it demonstrates a deep understanding of topics like ISR latencies, peripheral compatibility, memory and instruction requirements, and other stuff that really matters to engineers but is typically beyond the reach of a casual reader. As a result, news in the field is very, very rarely unbiased, informative reviews - there are a lot of 'new product' press releases unsubtly promoted as 'articles'.
An embedded systems engineer at some company building a product using these microcontrollers would not write up this article. Not only is the comprehensive knowledge required uncommon even among skilled engineers, it would take dozens of non-billable hours and would make it easier for competitors. And we typically have very narrow domain expertise, so the ability to create a website and rich blog experience like this would be unusual. (The database issues the site had under load this morning are excused!) Open-source and information sharing are not commonplace within embedded systems development.
And you won't be able to easily generate this comparison yourself without a lot of work. You might think that the numbers described would be front-and-center on the first page of each datasheet for these microcontrollers. But you'd be wrong, or at least naive. Each manufacturer's marketing department exaggerates their numbers in slightly different ways, so the first page of the datasheet is inevitably worthless bull shit.
> This long article requires in-depth knowledge of a large number of microcontrollers, and the time, desire, and ability to verify, collate, and summarize all that information.
So, Google needs to train its AI to recognize this instead of relying on PageRank.
We still need page rank to measure authority, though. Otherwise other AIs will soon mass produce content which looks like it "requires in-depth knowledge of a large number of microcontrollers, and the time, desire, and ability to verify, collate, and summarize all that information" but is actually nonsense. Alternatively, we trick spam producer into doing valuable work.
Google doesn't _need_ to recognize this because it's objective isn't to show you the best result but just a good enough result to keep you using Google.
A search engine that only gives you the best result wouldn't even need the search results page. And that's where Google likes to sell ads.
How about an (technical advisor to a) buyer for a company like Adafruit, that both resells SoC/microcontroller "developer boards", and puts out their own? I could see them writing up their experience in trying to find good SoCs to base their products on, and how they compare to the ones in the boards they resell.
Methinks most business owners/managers, especially MBA-types, would see that as giving away sensitive info/competitive advantage to competitors and nix the idea.
I know that for my own part, a lot of what I blog when I blog tends to be motivated either from increasing my professional visibility or because it's something I spend time doing anyway and where I figure I it'll be useful to someone (I know I've done it right whenever I find my own blog posts in search results a couple of years later and they answer my question)
And it works, depending on who you are trying to reach.
I know that this article has me completely convinced that Jay could be extremely effective assisting with any of my projects. Much more so than any Ivy League diplomas, years of experience or lack thereof, glossy brochures, testimonials, or other conventional marketing tools. And it's significantly more effective than any FizzBuzz quiz I could administer to potential candidates.
For all the foibles of business politics, expert opinion from within the organization is extremely persuasive when it's available (though an organization that really needs Jay's services because they have no internal experts would have a hard time using this article).
Also people like to talk about their interests and work. I don't see a lot of my own writings being about profit. My interest in them existed well before they helped me get into this profession.
> Jay is available for engineering services on a contract basis
I'm not saying that's the number one goal here, I think this very thorough and informative piece goes way beyond a personal advert. But it must help (and rightly so).
> I suppose the lessons are: there's room for lots of improvement in search engines, or SEO really works and you better promote yourself no matter how good you are, or word of mouth (Hacker News in this case) is still the way to find the best stuff.
The way I look at it: SEO really works, and shit sources tend to spend a lot on it, while quality sources tend to not spend anything on it at all - therefore a lot of quality stuff is hard to find, unless you can hear the right words from the right mouths.
(IMO SEO, beyond making your website not look like shit, is user-hostile behaviour. It games search engines into serving users links to those who spend effort on gaming search engines instead of links to content the user was looking for.)
Indeed. SEO how it's referred to and done most of the time is almost an anti-pattern: People adjust their system (website) to implementation details of another service (search engines) instead of just fulfilling the "public contract" (exposing good, accessible content).
But that's dangerous, as search engines evolve, change behavior and occasionally penalize this kind of behavior. That's why I always tried to explain to customers they should just focus on good content for humans and not care about explicit SEO after they claimed someone had told them that Google had recently "changed their algorithm" and now did XY.
There are SEOs that cannot be detected and at the end you have quality articles after hundreds of pages. This is the case for maibstream subjects like recipes where you compete with ten of thousands of SEOers.
Google is really good for general information, but really bad for domain specific information. Piglet (in contrast) is really good at domain specific, but can be poor in the general case (although I haven't seen that yet). We track discussions around the articles, as opposed to linked pages. This typically provides more information. You can read about it on our about page.
For instance, this article appears right at the top of the search results: https://imgur.com/a/AkUTb
I recommend reaching out to me if you're interested - we are currently doing a private beta. Spcifically, the application is for investing (we are building an AI financial advisor), but I personally use it as a search engine as well.
Is there any way to use just the search engine bit? I don't care much about investing, but it'd be great if you could take the investment-specific information out of the free/public account and just returned the results for searches.
I've noticed some _very_ odd things going on with Google Search results recently. The other day I tried a search for an error message from SQLAlchemy, the results were poor so I added the word `SQLAlchemy` to the query and it removed _all_ answers about databases and returned a bunch of stuff about mineral mining.
To get anything useful out of Google, I often have to resort to using "verbatim" results from the search tools. Those results are much closer to the old Google than the standard search, but not perfect.
I remember how revolutionary late 90s Google was in how it would show you exactly what you were looking for, or pages with relevant content, whereas the other popular search engines was just be a jumble of stuff.
It seems like in order to monetize search, Google has had to regress what once worked well.
Google is increasingly returning irrelevant results in the effort to deliver the targeted content their advertisers want. I predict a future for a subscription funded search engine that just delivers lexically relevant results without all the bullshit. I for one would gladly pay to use something better than Google. The advertising monetization model is broken.
Second your opinion, Google hasn't been crazy good at domain specific search for me lately.
Google video search is also weaker than a pure youtube search for me too, so much so that I only use youtube search nowadays which is perhaps their intent...
I've noticed recently (few years maybe) that Google "punishes" results from blogs and less known domains. Searching for some specific keyword combination often results in skipping first few pages of cruft from popular sites just to find exactly what I was looking for in some blog post buried >3 pages deep.
Sometime around the housing market crash (though unrelated, I assume) Google seemed to give up on fighting spam sites and just harshly down-ranked anything that's wasn't an ultra-high-traffic site. A little while later (2010? 2011?) their search got much "smarter" and it became difficult to impossible to trick it into giving you less-popular results using well-crafted searches. A big (and very good!) chunk of the web became way less visible over that span of 2-3 years or so—and my impressive-even-to-other-developers Shaolin Google-Fu was defeated by Google's Tiger Style :-(
You can also see the ugly results of their new ranking system any time you search for a simple answer that deserves maybe two sentences, and are greeted with ten search results all of which are a dozen rambling, content-free paragraphs, with a title closely matching your search and with your answer buried somewhere inside (maybe). That reads as "high value" and "high relevance" so google eats it up, even though it's 100% grade-A garbage content.
One other thing I've noticed, and it's been happening for a few years. Is that somehow, magically, there are websites out there that have the exact combination of my search results in their title. Almost as if Google forwarded my search to some party which then went ahead and created a dummy "proxy-search" site with the exact heading, and whose body just contains their own results for the same search criteria I used.
I've had this exact experience many times myself and I'm also at a loss to explain how it happens unless Google is (unintentionally) facilitating it. An observation I can add is that the dummy site may not appear on my first search for "very unlikely phrase". But if I keep tweaking my Google search, a dummy site that has "very unlikely phrase" as its title soon appears in the results.
It shouldn't be too hard to make Google index all 3 words combinations from your site. I've inadvertently made them index every city location at my country once.
iirc the original pagerank algorithm was based on weighting page value by how often other pages linked to it, with higher weighted pages lending greater to weight to outlinked pages. Obviously google's extended it greatly since, but that core behavior of lowering unpopular blogs in search results should always be there; in terms of linkage, they're not much different from spam/trash
It's been on my mind for a while that I could use a search engine that's seeded only with sites that contain little or no advertising or tracking.
A starting point could even be the corpus of HN articles and comments, perhaps augmented with entries from the Common Crawl archives and traditional crawling.
I tend to be interested in novel and creative work, and it seems to me that most of the sites I find most rewarding are not commercially driven.
Over the last 1.5 decade I've been seriously on-line I developed this heuristic: quality of content is inversely proportional to advertising, and also inversely proportional to page bloat.
"Search engines like Google are indispensable, able to find answers to all of your technical questions; but along the way, the fun of web surfing was lost. In the early days of the web, pages were made primarily by hobbyists, academics, and computer savvy people about subjects they were interested in. Later on, the web became saturated with commercial pages that overcrowded everything else. All the personalized websites are hidden among a pile of commercial pages. Google isn't great at finding those gems, its focus is on finding answers to technical questions, and it works well. But finding things you didn't know you wanted to know, which was the real joy of web surfing, no longer happens. In addition, many pages today are created using bloated scripts that add slick cosmetic features in order to mask the lack of content available on them. Those pages contribute to the blandness of today's web.
The wiby search engine is building a web of pages as it was in the earlier days of the internet."
Thanks - that post had been on my readar, but got forgotten.
The paragraph you quote pretty much sums up my feelings, though I'm not particularly concerned with finding old school sites.
Google used to do a good job of answering technical questions, as well as finding the little gems of originality (and lots of lovely mediocrity too, of course).
I don't think the two need be mutually exclusive. It's just that what I consider the good stuff is lost in the attention seeking nonsense that the web has become. Filter that out, and I have a feeling that lots of great stuff will become a lot more accessible.
wiby is wonderful, I had a play around and found pangs of nostalgia, and stuff that made me laugh out loud. But I think the criteria of lightness of a site, for example, the amount of JavaScript used, means it will be missing many great contemporary sites that happen to use the technology of today.
Google is really good for general information, but really bad for domain specific information. Piglet (in contrast) is really good at domain specific, but can be poor in the general case (although I haven't seen that yet).
For instance, this article appears right at the top of the search results: https://imgur.com/a/AkUTb
I recommend reaching out to me if you're interested - we are currently doing a private beta. Spcifically, the application is for investing (we are building an AI financial advisor), but I personally use it as a search engine as well.
I suppose the lessons are: there's room for lots of improvement in search engines
There's also the tried and true distribution method of newsletters and mailing lists specific to a certain industry.
If you've been in embedded a long time, you probably know about Jack Ganssle and his "Embedded Muse" newsletter, which has been going on for about 20 years now.
>>> I suppose the lessons are: there's room for lots of improvement in search engines, or SEO really works and you better promote yourself no matter how good you are, or word of mouth (Hacker News in this case) is still the way to find the best stuff.
As someone who runs a similar blog and is first page for hundreds of search terms, Google is fantastic at discovering niche content. You can't even being to fathom how good it is.
The article is already the 4th google result if you look for "best micro controllers under a dollar".
If anything, the title should be improved by the author. "The Amazing $1 Microcontroller" is a very poor description of the content and noone is going to google that.
> The article is already the 4th google result if you look for "best micro controllers under a dollar".
That worked because you're mentioning the "$1" from the article's title. Google is not interpreting "under a dollar" as "cheap". As proof, I searched for "best microcontrollers under £1", "best microcontrollers under 50 cents", and "best microcontrollers under $2". Jay Carlson's article wasn't found in any of those cases.
A more commonsense search of "best cheap microcontrollers" doesn't find the article even in the top 100 results.
> "The Amazing $1 Microcontroller" is a very poor description of the content and noone is going to google that.
I thought the title intriguing and pithy. You don't want a web where every title is SEO-optimzed, do you?
There are thousands upon thousands of micro controllers. The space is crowded. I'm not surprised that google doesn't rank an article that is poorly worded.
I find the title intriguing too... but on HN and with 100 upvotes already. It's pretty terrible in a search context. If I am looking for a comparison, I'm not clicking on something called "the amazing microcontroller".
That will change now that it's at the top of HN, but I agree - it's currently very obscure to search engines.
My guess is that it's a new site? Only a few articles, and Open Site Explorer reports a Domain Authority of 18, which is approximately equal to zero :)
However, if the author keeps writing these articles at this level of quality, chances are in a year it'll dominate relevant SERPs.
This suggests an interesting "golf" game: Given any web page, find the shortest set of keywords that will cause Google to return that page in the first page of results.
Should probably exclude some operations like "inurl" or simply searching the entire page url.
G+ communities really shine for this kind of thing. I suggest subscribing to the various communities that take your fancy and make a point of checking once or twice a week.
A good search engine would pick up sentiment from discussion threads to improve ranking. But then you have to find the good forums, another hard problem.
The lesson is put in more work. Don't expect Google or your Grandmother to make things easy. They both will keep telling you, that is their job in life.
A large part of the population has turned into a bunch of quitter because of it.
I tried "best microcontrollers reviewed", "best low-cost microcontrollers", "survey of microcontrollers", and a bunch of others. Nothing that Google finds in the first 10-20 hits is as good this resource, other than Wikipedia, though very good, isn't a technical review.
I suppose the lessons are: there's room for lots of improvement in search engines, or SEO really works and you better promote yourself no matter how good you are, or word of mouth (Hacker News in this case) is still the way to find the best stuff.