2. when typing a company name in the chrunchbase demo, the first few milliseconds while it's searching show "no results found" instead of a spinning wheel or something... again, minor thing but adds to "psychological trust effect" in the software
A suggested revision to your licensing blurb (proposed revision in italics):
"If your application restricts any of these freedoms, such as commercial or closed-source applications, then the GPL license is unsuitable and you must contact us to buy a commercial license at contact@linkurio.us."
This looks incredible, and I'd love to take a look - but we use Allegro. We've also been experimenting with Orient and Titan. Do you have any plans to support different graph DBs?
We are in talks with Allegro and OrientDB and have clearly planned to develop connectors for these databases.
As for TitanDB, since we will probably develop our OrientDB connector on top of it's Gremlin API [1], it'll be little work to write the TitanDB connector after that (they have a Gremlin API too).
To clear out a possible minor confusion for readers here - Titan's only API is Gremlin (though there also is a lower level Java API). You guys should also have a look at ArangoDB, which also supports Gremlin.
Linkurious with upcoming TinkerPop3/Gremlin will definitely be quite interesting to use. Keep up the good work.
The quote "They trust us" gives me a very negative response as I am reminded of Mark Zuckerberg's "They trust me. Dumb fucks" quote. http://en.wikiquote.org/wiki/Mark_Zuckerberg
Maybe rephrase that to something like "trusted by" or "our trusting clients"?
edit (after downvotes into grey): Sorry for trying to share a negative first impression with the idea of helping the landing page to become better. If you did not get the same association just ignore the comment...
On the homepage, your pricing link is buried in the footer. (I see it's in the subnav on internal pages). Maybe consider promoting that link to the main navigation?
Thanks for the suggestion! We have 2 different products, the second one being the developer toolkit (http://linkurio.us/toolkit/) so we ended up displaying the pricing on each different page... but we're still unsure about the clearest approach :/
What is unclear here is why you are selling what is and should remain free and open source software without even crediting the authors of what appears to be 90% of your toolkit product.
TL;DR the README (https://github.com/Linkurious/linkurious.js), huh? The authors of the core/Sigma are cited in the documentation and license. Plus it is released under GNU GPLv3, so I don't understand why you complain. If you don't need the extra features, you are free to go with the core only.
After creating Gephi 8 years ago I'm still a big fan of open source as a production model, but definitely not as an economic model. Companies need support and high-end features to speed up their projects and to limit risks; this is what our tookit is about, adding more than 30 plugins.
Notice that we are happy to give back to the core when the original authors agree. Say thanks to us every time you click on an edge! :)
Do you have a partial trial to try using it first, or require a one year commitment?
I do lots of instagram & linking data for bloggers. This looks interesting.
As a long time Gephi user (~4 years) this is awesome. Congrats.
I don't currently have a project that could benefit from this platform, and the price point is a bit out of a reach for a college student, but I am really looking forward to see how this pans out.
Neat! Really wish I could try it on my dump of the PGP public keyserver network. A visual map of the web of trust would be cool. Alas, it's just a volunteer project, so no funds :(
Technical side: we recommend to display up to 2000 nodes and edges. Laptops < 2 year old can display and layout graphs up to 4000 nodes and edges but with stability issues.
Cognitive side: we recommend to hide nodes and edges as soon as you don't need them. One cannot ask the same class of questions to graph visualizations of very different sizes, see slide 19 on http://www.slideshare.net/Cloud/sp1-exploratory-network-anal...
This is actually similar to the work of a company here in the Netherlands, Eindhoven (http://www.synerscope.com). Infovis on big data. From what I know is that they spent a lot of time on performance in the financial scenarios. The amount of data is huge, they process in (under a) second(s) using gpu but they ran into trouble some year ago when a customer used a different type of DB (different order of joins, unions and selects I believe). What is the performance on linkurio? When talking multi-million or billions of transactions, will it scale? How long for a query to complete?
The biggest dataset used by one of our users is a genetic graph of 240 millions of nodes and edges using a single server. Linkurious will take a few hours to index the complete dataset. From then, the search engine delivers instant results with autocomplete, fuzziness and advanced query options; graph exploration queries take less than a second to complete (sometimes a bit more depending on the web client). We are still working on improving our data indexing strategy to gain performance.
Synerscope has a strong approach to data analysis, and Danny Holten is well-known in the infovis community. I don't think that they provide a search engine thought, you have probably more information on their product.
39 comments
[ 3.1 ms ] story [ 69.3 ms ] threadWe've just launched and we need your feedback. Thanks! :)
Seb
Feedback items:
1. launching the demo triggered chrome's popup blocker (small conversion funnel disrupter... easy fix)
2. when typing a company name in the chrunchbase demo, the first few milliseconds while it's searching show "no results found" instead of a spinning wheel or something... again, minor thing but adds to "psychological trust effect" in the software
Looks awesome otherwise.
We currently only support Neo4j and have planned the development or a OrientDB connector (see my other comment about that).
Regarding the missing spinning wheel while searching, we are currently fixing that :)
"If your application restricts any of these freedoms, such as commercial or closed-source applications, then the GPL license is unsuitable and you must contact us to buy a commercial license at contact@linkurio.us."
Out or curiosity, what were the graph viz engines you know/use today?
We are in talks with Allegro and OrientDB and have clearly planned to develop connectors for these databases.
As for TitanDB, since we will probably develop our OrientDB connector on top of it's Gremlin API [1], it'll be little work to write the TitanDB connector after that (they have a Gremlin API too).
[1] http://orientdb.com/docs/last/Gremlin.html
We're also eager to see any binary protocole to communicate with graph databases.
Maybe rephrase that to something like "trusted by" or "our trusting clients"?
edit (after downvotes into grey): Sorry for trying to share a negative first impression with the idea of helping the landing page to become better. If you did not get the same association just ignore the comment...
After creating Gephi 8 years ago I'm still a big fan of open source as a production model, but definitely not as an economic model. Companies need support and high-end features to speed up their projects and to limit risks; this is what our tookit is about, adding more than 30 plugins.
Notice that we are happy to give back to the core when the original authors agree. Say thanks to us every time you click on an edge! :)
For the HN crowd, worth mentioning two other startups here:
1. KeyLines (Cambridge Intelligence) has great visualizations and core integrations (Neo4j, Titan)
2. For exploring bigger graphs (ex: entire live netflow topologies from Splunk) and graph analytics+ML (ex: Spark/GraphX), we're alpha testing our GPU-accelerated client/cloud approach. Email alpha@graphistry.com :)
I like a lot about:
-- Cola (flexible constraint-based): http://marvl.infotech.monash.edu/webcola/
-- Vivagraph (big graphs in webgl) doesn't get enough love: http://www.yasiv.com/graphs#Bai/rw496
Getting the best of both is hard, hence our real-time GPU clusters. It'll be longer before we can usefully open that up that layer =/
Still, GPU clusters beat all for performance. I'd love to see a demo :)
http://linkurio.us/toolkit/#pricing
Our Open Source toolkit (Linkurious.js) has examples in th github Wiki, which can be accessed through here: http://linkurio.us/toolkit/
We currently have no plans on integrating with graphviz.
For Linkurious.js there are some examples over here https://github.com/Linkurious/linkurious.js
All infos on the toolkit are available on http://linkurio.us/toolkit/
https://keyserver-elasticsearch.daylightpirates.org/keyserve...
http://linkurio.us/toolkit/#pricing
Cognitive side: we recommend to hide nodes and edges as soon as you don't need them. One cannot ask the same class of questions to graph visualizations of very different sizes, see slide 19 on http://www.slideshare.net/Cloud/sp1-exploratory-network-anal...
Synerscope has a strong approach to data analysis, and Danny Holten is well-known in the infovis community. I don't think that they provide a search engine thought, you have probably more information on their product.