Web Development - WAYWO - #8 - Developers

[QUOTE=Sharpshooter;43323548]So so true, who the hell would line it up based off the square perimeter of the circle? Am I a bad designer because I just move it until it looks right? I don't think so! Yes, use your eyes and your fingers and stop being so anal about a non-existant problem.[/QUOTE] Well it would be nice for a solution you can put in more than 1 place with different scales. If you want to scale it then your nudging might lead to issues. Having a nice solution you can just "drag and drop" into your project would be nifty.

[QUOTE=mdeceiver79;43324191]Well it would be nice for a solution you can put in more than 1 place with different scales. If you want to scale it then your nudging might lead to issues. Having a nice solution you can just "drag and drop" into your project would be nifty.[/QUOTE] How would scaling it cause issues? Things don't magically move to different places when you resize them. It literally takes 30 seconds to eyeball it and have it look better than what this guy did. It's just a goofy attempt to make something simple seem complex. It's a waste of time lol.

Installed the Base16 color scheme for Sublime Text last night, suddenly. Matrix. [IMG]http://i.imgur.com/KyaIXnN.png[/IMG] [editline]27th December 2013[/editline] I'm not actually using it, just came across it :v:

Anyone into building HTML5 Applications at the moment? I've just discovered [URL="http://xdk-software.intel.com/"]Intel XDK[/URL]. Seems like a really useful tool, and pretty much automates the building process for all major platforms. I've also got my hands on a Firefox OS device (ZTE Open) - it's no rival to iOS/Android at the moment, but it's completely open source, and seems pretty promising. Linux based at the core, but the whole frontend is built with web technologies.

I didn't read the article, just quickly skimmed it. But I can see where it might be usefull to know a few of these things. He isn't talking about pixel perfect positioning but more a general where should you position something. Eyeballing something isn't always a good thing. Some things are acctualy tested to find out what the best positioning is. Again, I didn't read the article completely so I don't know the exact situation, just saying eyeballing everything isn't the best idea. There is some logic behind positioning. Most of it is common sense but a few subtle things can make the difference between a user friendly application and a not so user friendly application.

There's a reason that fine sizing/positioning things like kerning and [url=http://en.wikipedia.org/wiki/Overshoot_(typography)]overshoot[/url] in typography are done by eye and not by some randomly applied math. The math doesn't give you the best result. Design is made for the human eye, so doesn't it make more sense to let your own human eyes be the final authority on whether something is "pixel perfect" or not? Literally the only thing he is doing in that article is positioning two icons so that they are visually centered in the quarter circle that contains them, and I don't think the method he chose gave him the best result he could have achieved. Of course if you're designing on a grid system or adding a border around something that you want to be a standard width it makes sense to measure it instead of eyeball it. Here is his mathematically "centered" version: [IMG]http://i.imgur.com/MTKgcQC.png[/IMG] Here is one I spent under 2 minutes eyeballing (the colors behind the icons aren't perfect since the dialog is a little transparent): [IMG]http://i.imgur.com/jnNdQp1.png[/IMG] Personally, I think the one I eyeballed feels more balanced.

Either way I heavily dislike the pie slice controls, I realize it was integral to his entire example, but they still look [I]terrible[/I] to me ( pref: rectanglular )

[QUOTE=KmartSqrl;43335807]-different cheeses-[/QUOTE] Personally, I don't see the difference to make me go "Oh yeah this one totally kicks the other's ass!!"...

[QUOTE=Coment;43336417]Personally, I don't see the difference to make me go "Oh yeah this one totally kicks the other's ass!!"...[/QUOTE] It's a subtle difference, but lots of small differences like this add up when you're looking at the project as a whole. The point I'm making is that doing it by eye is faster than picking a somewhat arbitrary mathematical rule to follow [I]and[/I] it ends up looking better.

Working on a project with a few others, essentially a hybrid between game banana, a forum and a few other things, while trying to stay unique in it's own right. There is a lot more to it than just that, but keeping it short and sweet until we're ready to release any real details (if it even gets any interest). Building it with PHP and MongoDB. Loving the ease of MongoDB and how you can just save and retrieve anything in any format, no hassle of tables (and awesomely fast). I've just finished writing the full module base, with a nice view system, and js wrapper to automatically handle all ajax requests and responses for you. This is all designed to make the development (both present and future) and maintenance of the site as easy and painless as possible. Only problem is making it look good, if there is anyone interested in joining the project that can help bring something visual to it, I am happy to share the google drive doc we have containing all the info/ideas we have for it thus far. We're using git (hosted using gitlab) to manage it all.

I don't know that I'd use mongo for something like a community site where you're going to have lots of documents that should be referencing other documents (like a comment needing to point at the user who left the comment). This article explains why fairly well, with a real world example: [url]http://www.sarahmei.com/blog/2013/11/11/why-you-should-never-use-mongodb/[/url] If you want the document style database with waaaay better support for relationships between documents, check out [url=http://www.orientdb.org/]OrientDB[/url].

[QUOTE=KmartSqrl;43337172]I don't know that I'd use mongo for something like a community site where you're going to have lots of documents that should be referencing other documents (like a comment needing to point at the user who left the comment). This article explains why fairly well, with a real world example: [url]http://www.sarahmei.com/blog/2013/11/11/why-you-should-never-use-mongodb/[/url] If you want the document style database with waaaay better support for relationships between documents, check out [url=http://www.orientdb.org/]OrientDB[/url].[/QUOTE] Just skim read it, and I can see where you are coming from, my main question is, is OrientDB as easy to implement? Can I just throw it a PHP object and let it deal with it? I am very eager to not get bogged down by database structures, writing big joins, etc etc which I am so used to with MySQL. Currently we have barely used anything to do with the database, we have only just reached that level (thus far been building as much as possible to make building the site and the modules fast, building the actual modules is where the db comes in). So while swapping the DB engine now is an option, it's one i'd like to be sure is going to do what we want, and how we want. Currently Mongo does, however like that article points out, it has drawbacks.

Are you using some kind of ORM? Looks like there's an in-progress orient adapter for doctrine here that is being actively developed: [url]https://github.com/doctrine/orientdb-odm[/url]

[QUOTE=loony383;43336957]Only problem is making it look good, if there is anyone interested in joining the project that can help bring something visual to it, I am happy to share the google drive doc we have containing all the info/ideas we have for it thus far. We're using git (hosted using gitlab) to manage it all.[/QUOTE] I might be able to help on the development and some of the design. Can I see the doc so I can decide if it's a project I'd like to join?

I'm getting sooooooo close to finally being able to launch this ground-up rewrite of Spoonfed Style. Getting really excited so I'm having a hard time forcing myself to keep working slow and methodically instead of just messily bashing out the last little group of things I have to do still haha.

[QUOTE=KmartSqrl;43338491]Are you using some kind of ORM? Looks like there's an in-progress orient adapter for doctrine here that is being actively developed: [url]https://github.com/doctrine/orientdb-odm[/url][/QUOTE] Currently just using the direct PDO driver for MongoDb and saving PHP objects in directly. As a result of MongoDB's syntax, we just pass it an array of options on retrieving. Though currently we only actually use this in 2 places so to swap out to an alternative wouldn't be difficult at all. The companies I have worked for have only ever required me to use MySQL, either on it's own or with a wrapper around it, so anything outside of that is still a learning experience for me (though one I am happy to embrace, to use something other than MySQL is refreshing). [QUOTE=supersnail11;43338773]I might be able to help on the development and some of the design. Can I see the doc so I can decide if it's a project I'd like to join?[/QUOTE] Absolutely, just drop me a pm with your e-mail and I'll share it with you.

Updated [url=http://96.47.239.164/odessey/]Odessey[/url] to version 1.1, which includes a few changes, like updated packages. You can download it from [url=http://96.47.239.164:8080/cat/odyssey-v1.1.ova]here[/url]. On a side note, I couldn't pay for the domain anymore, so now I have to use the IP :(

Decided that for now at last (may change in the future) to stick with MongoDB. If we ever get large enough that the drawbacks of Mongo are an issue, i'm sure we can invest the time to move over to another DB engine. For now, gonna stay with mongo because as far as I can see, generally more knowledge available about it (plus I know others experienced in it should I get stuck). Will not rule out moving to orient or any other engine, just not sure that it is a a priority right now as the project is still early days and more info is available for mongo.

Let's go! [thumb]http://puu.sh/62i0i.png[/thumb] such design. much flawless. find an error in this exquisite design, I dare you

[QUOTE=P1raten;43343917]find an error in this exquisite design, I dare you[/QUOTE] The anti-aliasing on the title is kinda gross.

Just finished writing an class system to manage html forms, even down to submission and completion. You just create a class that extends form, pass the parent constructor an array of fields, and set any options of those fields (such as validation requirements, default values, and anything else), and it will generate the full html required for it, including handling the ajax submission and validation of the form. For a site that relies heavily on user input, this should really save a lot of time.

[QUOTE=loony383;43343514]Decided that for now at last (may change in the future) to stick with MongoDB. If we ever get large enough that the drawbacks of Mongo are an issue, i'm sure we can invest the time to move over to another DB engine. For now, gonna stay with mongo because as far as I can see, generally more knowledge available about it (plus I know others experienced in it should I get stuck). Will not rule out moving to orient or any other engine, just not sure that it is a a priority right now as the project is still early days and more info is available for mongo.[/QUOTE] I used to use MongoDB until a few weeks ago. I did some more research lately, and it turns out, Mongo has some terrible issues underneath [[URL="http://blog.engineering.kiip.me/post/20988881092/a-year-with-mongodb"]1[/URL], [URL="http://nyeggen.com/blog/2013/10/18/the-genius-and-folly-of-mongodb/"]2[/URL], [URL="http://hackingdistributed.com/2013/01/29/mongo-ft/"]3[/URL]], and most large websites that were using it have moved on. [[URL="http://blog.engineering.kiip.me/post/20988881092/a-year-with-mongodb"]1[/URL], [URL="http://mcfunley.com/why-mongodb-never-worked-out-at-etsy"]4[/URL], [URL="http://svs.io/post/31724990463/why-i-migrated-away-from-mongodb"]5[/URL], [URL="http://nosql.mypopescu.com/post/30830925905/from-mongodb-to-riak-at-shareaholic"]6[/URL], [URL="http://basho.com/customer-io-gains-6x-speed-improvement-by-moving-from-mongodb-to-riak/"]7[/URL]] Keeping in mind that your data isn't really suited to MongoDB in the first place, it might be a good idea to take a look at your options and figure out which one works best (remember - there's a time and a place for RDBMSes)

[QUOTE=Ac!dL3ak;43342766]Updated [url=http://96.47.239.164/odessey/]Odessey[/url] to version 1.1, which includes a few changes, like updated packages. You can download it from [url=http://96.47.239.164:8080/cat/odyssey-v1.1.ova]here[/url]. On a side note, I couldn't pay for the domain anymore, so now I have to use the IP :([/QUOTE] [del]creating a mirror right now[/del] [url=http://IOException.at/odyssey-v1.1.ova]Done[/url] What domain expired?

[QUOTE=supersnail11;43348872]I used to use MongoDB until a few weeks ago. I did some more research lately, and it turns out, Mongo has some terrible issues underneath [[URL="http://blog.engineering.kiip.me/post/20988881092/a-year-with-mongodb"]1[/URL], [URL="http://nyeggen.com/blog/2013/10/18/the-genius-and-folly-of-mongodb/"]2[/URL], [URL="http://hackingdistributed.com/2013/01/29/mongo-ft/"]3[/URL]], and most large websites that were using it have moved on. [[URL="http://blog.engineering.kiip.me/post/20988881092/a-year-with-mongodb"]1[/URL], [URL="http://mcfunley.com/why-mongodb-never-worked-out-at-etsy"]4[/URL], [URL="http://svs.io/post/31724990463/why-i-migrated-away-from-mongodb"]5[/URL], [URL="http://nosql.mypopescu.com/post/30830925905/from-mongodb-to-riak-at-shareaholic"]6[/URL], [URL="http://basho.com/customer-io-gains-6x-speed-improvement-by-moving-from-mongodb-to-riak/"]7[/URL]] Keeping in mind that your data isn't really suited to MongoDB in the first place, it might be a good idea to take a look at your options and figure out which one works best (remember - there's a time and a place for RDBMSes)[/QUOTE] Still reading through, and those are some eye-opening examples of why Mongo is the wrong solution. So now I ask, which is the right solution? I would like to stick to a engine that has no schemas, and allows me to save my php objects with as much ease as possible, generally as I can see this changing quite drastically over time and I don't want to be battling and slowed down by having to update schemas and all that nonsense which I am fed up with in MySQL. (Don't get me wrong, I do like MySQL, but I feel that on this project, it would be more of a problem and isn't really suited due to the vast variety that there could be in the data being stored). I've started to look at Riak and really like what I see, would this be a viable alternative without the drawbacks?

[QUOTE=loony383;43348988]Still reading through, and those are some eye-opening examples of why Mongo is the wrong solution. So now I ask, which is the right solution? I would like to stick to a engine that has no schemas, and allows me to save my php objects with as much ease as possible, generally as I can see this changing quite drastically over time and I don't want to be battling and slowed down by having to update schemas and all that nonsense which I am fed up with in MySQL. (Don't get me wrong, I do like MySQL, but I feel that on this project, it would be more of a problem and isn't really suited due to the vast variety that there could be in the data being stored). I've started to look at Riak and really like what I see, would this be a viable alternative without the drawbacks?[/QUOTE] It completely depends on the project. For what you're doing, you have a lot of relationships - users to threads/content, users to posts, users to ratings, ratings to posts, ratings to content, tags to posts, etc. - so a database that handles relationships well would be a good choice. Riak sacrifices consistency for availability and partition tolerance (AP database) - so while data may not be consistent across nodes, it's going to be fast and it's going to be available. This is great in an environment where you have a lot of reads, and they all need to be fast - session stores, URL shorteners, etc ([URL="http://www.slideshare.net/argv0/riak-use-cases-dissecting-the-solutions-to-hard-problems"]here[/URL]'s some [URL="http://www.infoq.com/news/2011/12/andy-gross-riak-database"]stuff[/URL] about that). That's not to say that you shouldn't use Riak - you can use multiple databases. Perhaps a relational database as a backend, and Riak or Redis as a caching layer, or using a document store for your content and Riak for your session store. You could use Riak as your main database ([URL="http://basho.com/shopzilla-selects-riak-for-core-data-services/"]it's worked for[/url] [url=http://java.dzone.com/articles/riak-production-voxer]some people[/URL]), but you might also benefit from the complex queries that SQL can provide.

I turned this (a list of the songs in my Spotify playlist): [code] http://open.spotify.com/track/24lMahinQk059H6rUEqWF5 http://open.spotify.com/track/273MQ5Sx0fvCIGTS2lpRnX http://open.spotify.com/track/0LirIy8fWL655xLgPAxp3i http://open.spotify.com/track/3Dn0XMcTopXeWGdUn3KZNB http://open.spotify.com/track/20k9otzKdPr3wG9lU1w0kj http://open.spotify.com/track/4QN6uGhhQb8ZCIhoUuIl1x http://open.spotify.com/track/49vJHbrjKkRMpUeEGI1TEB http://open.spotify.com/track/5qIVeybeQ1K1IUGyk9r0sz http://open.spotify.com/track/0FubbOyFuWp0hAq6V8a1AI http://open.spotify.com/track/3EXCJvehlOGEP9ntHaEMAg http://open.spotify.com/track/5OmBvjj68715ECVXG4lsSY http://open.spotify.com/track/4l0EXL8Vr5hHk5JAMQlDx6 http://open.spotify.com/track/0d28khcov6AiegSCpG5TuT http://open.spotify.com/track/5ADMEkN23dk9CuEC5uDN53 http://open.spotify.com/track/7v66Ds7gmm0lM2nF0Ih0GT http://open.spotify.com/track/7252KLX6Osc0A02pjINRfJ http://open.spotify.com/track/4dNCg774LzE2kmtbMOLdP7 http://open.spotify.com/track/4kO7mrAPfqIrsKwUOK5BFx ... 434 more [/code] Into this: [URL]https://gist.github.com/gabrielecirulli/04b86a68bcd6a2eaed35[/URL], which is a list of the artists I'm most likely to like based on those artists, as suggested by Tastekid. I basically wrote 4 ruby scripts which [URL="https://gist.github.com/gabrielecirulli/551f9c14fba56eb26be0"]fetch the artist names[/URL] from the Spotify links (by downloading the pages and extracting it from the page title), query Tastekid for [URL="https://gist.github.com/gabrielecirulli/873ef68fd4a382575220"]recommendations for each artist[/URL] (avoiding duplicates), parse the list outputted by the previous script into [URL="https://gist.github.com/gabrielecirulli/038e4081bb568eaf54de"]a JSON structure[/URL], and aggregate the data to give me a list that looks like this, where the most suggested artists are at the top: [code]26 Surkin 20 Yuksek 18 Sebastian 17 Para One 17 Teenage Bad Girl 16 Pedro Winter 16 Data 15 Autokratz 15 Les Petits Pilous 15 DJ Mehdi 14 Feadz 14 Anoraak 12 Danger 12 Russ Chimes 11 Grafton Primary 11 Vitalic 10 Midnight Juggernauts 10 Fred Falke 10 Cut Copy 10 The Outrunners 10 Simian Mobile Disco 9 Kavinsky 9 Lifelike 9 College 9 Minitel Rose 8 Classixx 8 The Toxic Avenger 8 Tommy Trash 8 Don Rimini 8 Krazy Baldhead 8 MSTRKRFT 7 Bingo Players 7 Dada Life 7 Alan Braxe 7 Porter Robinson 7 Viceroy 7 Amtrac 7 Bobmo 7 The Golden Filter 7 Xinobi 7 Miami Horror 7 Fear Of Tigers 7 Dillon Francis ... about 1000 more [/code] Turns out I don't like any of the most suggested artists :v:

I'm sad you don't like Surkin.

[QUOTE=TerabyteS_;43350564]Spotify stuff[/QUOTE] This reminds me - does anyone know a good way to track statistics for Spotify? Or even just an API for getting a user's latest plays. I love using the Last.fm API but I don't really feel comfortable investing it when it's probably going to get shutdown by November 2014 latest.

[QUOTE=Shadow801;43351837]This reminds me - does anyone know a good way to track statistics for Spotify? Or even just an API for getting a user's latest plays. I love using the Last.fm API but I don't really feel comfortable investing it when it's probably going to get shutdown by November 2014 latest.[/QUOTE] So why invest in Spotify? They've been loosing money every year and will most likely also shut down in a year or so. -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- [thumb]http://puu.sh/63kca.jpg[/thumb] I'm not entirely sure if this looks good or not. Would love some feedback. EDIT: Debating whether or not I should put a content container in the middle of the wallpaper and make the width < half or something along those lines.

[QUOTE=supersnail11;43349872]It completely depends on the project. For what you're doing, you have a lot of relationships - users to threads/content, users to posts, users to ratings, ratings to posts, ratings to content, tags to posts, etc. - so a database that handles relationships well would be a good choice. Riak sacrifices consistency for availability and partition tolerance (AP database) - so while data may not be consistent across nodes, it's going to be fast and it's going to be available. This is great in an environment where you have a lot of reads, and they all need to be fast - session stores, URL shorteners, etc ([URL="http://www.slideshare.net/argv0/riak-use-cases-dissecting-the-solutions-to-hard-problems"]here[/URL]'s some [URL="http://www.infoq.com/news/2011/12/andy-gross-riak-database"]stuff[/URL] about that). That's not to say that you shouldn't use Riak - you can use multiple databases. Perhaps a relational database as a backend, and Riak or Redis as a caching layer, or using a document store for your content and Riak for your session store. You could use Riak as your main database ([URL="http://basho.com/shopzilla-selects-riak-for-core-data-services/"]it's worked for[/url] [url=http://java.dzone.com/articles/riak-production-voxer]some people[/URL]), but you might also benefit from the complex queries that SQL can provide.[/QUOTE] Haven't got time to read those links right now but will do later. Based purely on what you have just said, the chances of outgrowing a single node right now are pretty tiny. So thinking of going with Riak (The speed sounds awesome) and if we ever need to deploy to multiple database servers, rather than deploying another rake, could deploy another engine and migrate some of it to that, depending on what data suits which. That is likely to be far enough in the future that we could make a much more informed and educated decision. However it sounds like Riak is a good option until that time comes.