|
enhzflep wrote: I wrote the reader program while learning about threading, socket programming and custom controls.
Very cool!
Marc
|
|
|
|
|
But how would you filter it?
Keywords wouldn't be much use, because they would restrict you to a few topics (half of which you're probably not interested in), and would have to be updated/added to so much that they'd end up filtering nothing out.
Theoretically, a backprop routine could be trained to provide you with lists of pages that would be of interest, but that would probably take so long to train that your interests would change three times before it was finished.
I can't really see a locally-installed app being able to deliver "pages that will be of interest to Markie" (or perhaps "Marcie", in your case), so maybe it would have to be down to some on-line giant to deliver pages-that-might-be-of-interest.
But, to avoid being bombarded with sites that pay the on-line giant, just use the newspaper method, and only "buy" the news sites that you like/trust/enjoy reading, or use some kind of crowd-sourcing/social-sharing/message-board solution, and only visit pages recommended by other individuals involved in the solution.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Mark_Wallace wrote: But how would you filter it?
NLP (Natural Language Processing). Extracts the semantic meaning of the content. I'm putting together an article on that at the moment.
Mark_Wallace wrote: Keywords wouldn't be much use, because they would restrict you to a few topics (half of which you're probably not interested in), and would have to be updated/added to so much that they'd end up filtering nothing out.
True, and even with NLP, one would have to set up triggers of entities, concepts, etc.
Mark_Wallace wrote: I can't really see a locally-installed app being able to deliver "pages that will be of interest to Markie" (or perhaps "Marcie", in your case), so maybe it would have to be down to some on-line giant to deliver pages-that-might-be-of-interest.
Not necessarily -- give it some RSS feeds, have the app know how to read through Facebook/Twitter/whatever, etc.
Marc
|
|
|
|
|
Marc Clifton wrote: NLP (Natural Language Processing). Extracts the semantic meaning of the content. I'm putting together an article on that at the moment.
I wanna read it!
Jeremy Falcon
|
|
|
|
|
Jeremy Falcon wrote: I wanna read it!
I have a preliminary version comparing three NLP services here[^], but keep in mind it's preliminary -- I'm getting a lot of good feedback from each provider that I need to incorporate, which will resolve some of the "odd behaviors" I mention in the article at the end.
Marc
|
|
|
|
|
Marc Clifton wrote: I have a preliminary version
I won't tell anyone.
Jeremy Falcon
|
|
|
|
|
Jeremy Falcon wrote: I won't tell anyone.
Even if they find out, it doesn't matter. The repository is there so that AlchemyAPI, OpenCalais, and Semantria can give me feedback on how poorly I'm representing them. So far, Semantria is proving the most difficult to work with with regards to their API. Stuff is NOT clear. But check out their pricing. Who would want to touch that anyways?
Marc
|
|
|
|
|
Marc Clifton wrote: But check out their pricing. Jeeze, those are laundering-drug-money prices.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Marc Clifton wrote: Who would want to touch that anyways?
When they go out of business, they'll figure it out.
Starting doing some recon on this stuff man and it ties in perfectly with "Web 3.0". This stuff is crazy awesome.
Jeremy Falcon
|
|
|
|
|
Jeremy Falcon wrote: and it ties in perfectly with "Web 3.0". This stuff is crazy awesome.
Indeed it does and is!
Marc
|
|
|
|
|
Marc Clifton wrote:
NLP (Natural Language Processing). Extracts the semantic meaning of the content. I'm putting together an article on that at the moment.
Information on NLP you may find useless useful[^]
I couldn't help myself.
Once you lose your pride the rest is easy.
I would agree with you but then we both would be wrong.
The report of my death was an exaggeration - Mark Twain
Simply Elegant Designs JimmyRopes Designs
I'm on-line therefore I am.
JimmyRopes
|
|
|
|
|
Marc Clifton wrote: NLP (Natural Language Processing). Extracts the semantic meaning of the content. I'm putting together an article on that at the moment. I look forward to that one. Lots.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Mark_Wallace wrote: I can't really see a locally-installed app being able to deliver "pages that will be of interest to Markie" (or perhaps "Marcie", in your case), so maybe it would have to be down to some on-line giant to deliver pages-that-might-be-of-interest.
It can built without much difficulties (see my article on "query intelligence ..., etc.") for us if you want such a meta level filtering (+ keywords). It's just whether or not this is a justifiable effort (it does take a few days away at least and they do add up), given so much other "more important" things to handle now.
|
|
|
|
|
No special software, but a list of sites that I browse over morning coffee and again in the middle of the day. For analysis, I read the Economist every week. (Not perfect, but useful).
/ravi
|
|
|
|
|
If you would like to improve it, ban contentless sites. Or aggregators. My time isn't wasted reading sites or purusing search links. My time is wasted with click bait and blog posts that summarize, summaries of blog posts taking 5 clicks to get to the real author.
Sadly, even CP is guilty of this. Sometimes the news in the news isn't the link but a link to some aggregator that does the link. And, AFAIK, CP has a full-time employee doing this! (Admittedly, it doesn't happen often so don't think I am calling out CP)
Do you want great news filtered just for you ... pay someone. A web site would probably charge a few hundred a month for the privilege or you could just hire an intern at $10/hr to constantly give you good links : )
|
|
|
|
|
Feedly is perpetually open on all my screens (about 200 feeds), and updated whenever I need a hit.
When I'm scanning for news items, I tend to open all the 'usual suspects' in their own tabs. They get updated a couple of times a day, depending on how much of a panic I'm in for news items. I don't use them in Feedly as they tend to be way to noisy.
By noon, I have Hacker News, Reddit/Programming and Reddit/Technology open and updating every 15 minutes (again, depending on what a panic I'm in).
Dark ages? Maybe. I prefer to think of it as more like the industrial revolution: there are some labour savers, but on the whole it's a dirty, smelly business.
TTFN - Kent
|
|
|
|
|
Kent Sharkey wrote: When I'm scanning for news items, I tend to open all the 'usual suspects' in their own tabs. They get updated a couple of times a day, depending on how much of a panic I'm in for news items. I don't use them in Feedly as they tend to be way to noisy.
Hmmm, you may be a good guinea pig for what I have in mind. Stay tuned.
Marc
|
|
|
|
|
I have a HTML file with list of my favorite links which is start page of my chrome. Occasionally I update that HTML file with new links.
In CP, I visit Insider, Soapbox & Lounge regularly. And GIT too where you could find Nish mostly.
Using this way, I save the typing time. And importantly searching & thinking time about sites as already I have those in my HTML file.
Please inform me if you find any software for this to save more time on this.
|
|
|
|
|
I really like Pearltrees.com. It allows me to quickly save links in categories for later browsing. I made one category for interesting news sites. I have occasionally tried your method, but I have too many browsers on too many computers. I have in the past put the links on a public page on the Internet, but it is always a nuisance to update. I have since installed Pearltrees on most of my browsers on all of my computers, and can see all of the links immediately on any other browser. The links are public (I use an anonymous alias). They even have a collaboration feature, in which a person may invite another to collaborate on a category. I have collaborated with strangers from all over the world on some.
|
|
|
|
|
Marc Clifton wrote: when it comes to using computers to filter out the crap
If that is your goal, I suggest printing it* out and using it** as toilet paper
*: the internet, that is
**: the printout
GOTOs are a bit like wire coat hangers: they tend to breed in the darkness, such that where there once were few, eventually there are many, and the program's architecture collapses beneath them. (Fran Poretto)
|
|
|
|
|
Multi-pronged, in order:
1. When I see a newsy website I like, I sign up for email notifications of new items. So, I use my email as the first source.
2. news.google.com is my home page.
3. Facebook for news from friends and relatives and sometimes other news that Facebook selects for me to look at.
4. Pearltrees - a browser add-on that I've installed on all of my browsers on all of my computers, allows me to quickly save links in categories and see them from any other browser or computer. I have one category for interesting news sites.
5. Click on the ads for news that I happen to notice in the margins.
|
|
|
|
|
DAILY: codeproject, geekstuff, makeuseof, sitepoint WEEKLY: Stackoverflow, Superuser AS NEEDED: Several Journals, BLOGs and tech support discussion sites. Hmmm... I guess I need to get a life or learn to hate code and problem solving!
Facebook noway!
"Courtesy is the product of a mature, disciplined mind ... ridicule is lack of the same - DPM"
|
|
|
|
|
|
Marc Clifton wrote: What I'm getting at is, it seems like we're still in the stone ages when it comes to using computers to filter out the crap and alert us to when something that we have said we're actually interested in occurs
When AI reaches the level that it can correctly assess my mood to guess what I might want to read then I am going to be much more excited about the real robots running around (since they won't require that level of AI.)
Until then I will just have to continue to randomly and impulsively bumble throughout the day finding interesting stuff to read.
|
|
|
|
|
jschell wrote: When AI reaches the level that it can correctly assess my mood to guess what I might want to read then I am going to be much more excited about the real robots running around (since they won't require that level of AI.)
Agreed, but that's not the goal.
jschell wrote: Until then I will just have to continue to randomly and impulsively bumble throughout the day finding interesting stuff to read.
The goal would be to provide you with more information than say, just the title of a post, to make your bumbling more efficient.
Marc
|
|
|
|