Take away from Mashup University and Mashup Camp

Here are some of my digested thoughts on Mashup Camp/Univesity and by extension, some of what is happening in the mashup space.

The tech and the market place are maturing and there is real money in the equation now. There are now some of the more traditional players in this space (IBM and Intel) and people are not focused on just consumer apps. There is work on making mashups happen for B2B and B2E and intranet enterprise. Rather than mashups IBM is calling them situational apps – built quickly for the particular situation. I am not sure I like the terms since it sounds too much like you are just building apps for Homeland Security or DoD. That aside I do think it is time for a better term than “mashup”.

Not all of this work is ad-revenue focused either, companies like StrikeIron are selling data to be used in custom mashups. They are also using attractive pricing so that groups within enterprises or small businesses can buy the data to join with data they have internally. Combine StrikeIron’s services with something like SnapLogic(open Source), OpenKapow, DAMIO and QUEDWiki (from IBM), or Dapper and you can enable non-IT people to start creating interesting applications that answer questions they couldn’t answer otherwise.

There are quite a few groups creating platforms for “building and hosting” mashups with a fancy UI. Very much targeted at end users without programming experience. Players include the usual popfly and google mashup generator, but names you might not have heard of include Bungee Labs, Zude, and Proto. Proto is interesting to me since it allows for graphing and statistics. I think the problem for the non-google and MS people is how to get their tech into end-users hands and get adoption. There was some great technology demonstrated but the question becomes how to get people using it.

For us more traditional developers I was very excited about more platform type technology, such as google gears. There has been a lot of talk recently about how mutli-threaded programming is the wave of the future. I think for web developers one of the waves of the future will be understanding how to write sporadically disconnected apps. I know that the web is stateless by default so it shouldn’t care about connectedness, but I am talking about making your web app as functional as possible as long as it can connect occasionally. For mapping folks this could totally be the new hotness.

Interesting note for those people dissing SOA and ESB – be thankful for those people because what they have basically done is built the infrastructure to allow this mashups. [UPDATE: sgilles points out the ambiguity in my statement. I meant that  SOA and ESB have created the intranet infrastructure to allow for enterprise mashups. The software now comes with features to allow easier sharing of enterprise resources internally. ] By putting data on the bus they have allowed the sharing of data and they have convinced enterprise app vendors to build it into the software. I am not sure who the guy was but he was right on when he said that this next round of enterprise apps upgrades will make the enterprise very mashable in about 1-2 years.

There is also another group who are trying to allow people to create and manipulate data in a Mashupable way. The simplest in this group would be Yahoo! pipes going all the way up to groups like apatar (Open Source), openkapow, snaplogic, and Dapper. As marcp states “could screen scraping be any easier”. I gotta get playing with this tech.

The speed geeking exposed me to a slew of awesome mashups. It was really hard to decide who should get my wooden nickel. I am not going to list them here but it was great to see how creative people could be in such a short time. Some were just toy applications to show a proof of concept while some were things that could have gone to market right then and there.
On the logistics and such – what a great event – kudos to David and his crew. The food was good, AOL brought Guitar Hero, there was a free espresso bar, and they had some awesome Mashup DJs (found Party Ben while writing this and he is even better) playing tunes on the first night (a little different than lounge singers, though some disagree). Another thing was that the people were really friendly and excited about showing all the cool stuff they were doing.

My final thought is that mashups are here and going to be growing – if you want to get on the bus start thinking of yourself as an enabler rather than a controller.

For more write-ups check out ProgrammableWeb…

Had a great time but I am exhausted

After 4 days at mashup-o-rama I am both exhausted and exhilarated at the same time. I am having trouble putting together full sentences at work today so I figure I should just wait until Monday to post my summary of the event. Short blurb for now – Excellent event – if you have a chance to go to one you should definitely attend.

Speed Geeking and afternoon sessions

Speed geeking was cool – a lot of creativity in the mashup space, image search, finding info on the web on the people in your address, doing routing for bars within 5 miles of your location so you can stumble around without losing your buzz.

Now at the intel Moblin

It’s not a laptop, it’s not phone it is that between. They don’t have the prototype here so no fun playing around with a geek toy. More powerful than a PDA bc it is based off a x86 instruction set. A reference platform currently for OEMs to build their own custom devices. They could build things like medical monitors, personal tv players, or a gaming platform. Here is there more technical spiel. I am not really a fan of these tweeners EXCEPT for the fact that it is x86 so write for a desktop or this device.

And so Mashup Camp begins

Just about to get started with the deciding on sessions. I really like this “spontaneous” creation of sessions. I will give you a list of some of the things I am thinking about attending. Perhaps the presence session, mobile UI, mashing up google checkout,spreadsheets, and map, another on licensing, mobile GPS, moblin from Intel – mobile mashups, API best practices, and I am leading a session on Amateur vs. Pro at 3-4 pm today. Off to the sessions…

Presence session

What is difference between presence and status – difference between phone world view versus WAP view – on the verge of beating up providers for their model for allowing people to play with their stuff.
Types of ways of thinking about users -

also about security from information

• Privacy
• Could be state dependent – work hours let coworkers see your presence , after hours just your friends
• filter and the providers

and the metadata about types of presence

• synicing it
• geotagging

Interesting point about open vs closed system and what could be made possible given the model. Lots of interesting issues – the discussion is a bit rambly. Seems like there is a lot of movement in this space. Should be fun to watch the tension between carriers walled garden and IP telephony…

Challenges for business use of mashups

• Came in late – But here are interesting ideas and thoughts – [thought] indicates idea from me that I share and are discussed in the room, everything else is from the smart people in the room. Just don’t want people to think all these things are my ideas or thoughts.
• There is confusion over licensing – You can use data within your company or just for consumption only but no storing. One guy states he thinks data in XML is no different from MP3 or music and it should be treated that way.
• Someone else states – Copyright was created to allow the spread of content and it is actually inhibiting for mashup.
• Rules actually help people bc it helps to clear up risk. Traditional IT is based off a rule of scarcity but mashup is a rule of abundance – how do you manage that value change.
• [thought] some of it related to IT seeing themselves as safe enablers of data versus protecters and lock downers
• A lot of services use ad-driven revenue and this is hard model for giving data away
• [thought] that there is a cultural problem – most people who interact with their computer are used to working with Excel and Word – and they don’t think of mixing and matching data sources on the fly. Interesting followup discussion. There is a plugin for excel that does this now.
• Some of the upgrades going on now in large IT are basically going to enable mashupability but not until 2009. So SMBs are the target are now but look to large IT a couple years out.
• [thought] Mashups can be seen as a good way to prototype what is a popular business value app that IT can then standardize and make 24×7. Allows users to show IT what they need in a more powerful way than a written use case.
• Firewalls and lack of interoperability.

More notes will be up on the wiki. Great session on the interface between business using mashups internally…

Lunch

Popfly from MS

Closed alpha – it is going to be all demos – no slides. Popfly is a consumer of windows live services or other services – meant for non-developers.

Running on silverlight so it only runs on those supported browsers. Coms pre-configured with blocks that provide services. Very fancy UI doing some fun effects. Very quick and easy to take a search and make it a slideshow. About 100 blocks are available out of the box with popfly but there is a mechanism for making your own.

Next demo – mapping twitter users – 10 latest posts. Wire together services like pipes but it puts a very nice front end on top of it.  Very nice, you can edit the javascript directly to customize a block (service) and it does intellisense to help complete the javascript.  Creating a new block (service) requires writing javascript to interface with the data source and some xml that describes your block.

Once you save you can either embed it, put it on live, on facebook or plenty of other places. But they are really trying to create a community of people with them sharing mashups that friends can rate and such.

There is a plugin for VS2005 that allows you to edit popfly projects and then upload them and share them with others.

Interesting tech – another in the GUI for making mashups and repurposing data from the internet. They have some basic scraping and they understand most feeds. Should be interesting to watch Goog and MSFTgo head to head on this one.

AOL in the afternoon

You know I didn’t AOL was doing anything like this at all – like xDrive or OpenAuth or an open API to AIM or an API to Userplane. They are doing a deep dive on openAuth with some interesting facts. I am not really interested in too many deep dives right now though so I am going to go save my brain cells – plus the network is slow so I am going to go take a break. I may give a smattering of AOL stuff throughout the afternoon.

AOL supports OpenAuth and OpenID – they are an idenity provider for openID but they do not accept OpenID for authentication which makes sense since OpenID just says this users owns a particular URL.

Open AIM stuff  – cool demos.
Change your status  in twitter and it changes your status in AIM and vice versa. Pretty seamless and a nice idea
Next demo even cooler – real time translator – using the google translation API – he types in english and the recipient receives it in spanish. He does it by writing a .NET plugin for AIM.

Time for xDrive with Lucas – for personal storage of media – like a bank for data, your data is there when you need it. Anything on the network should be able to see the data so they are using open standards  -dev doc shows using JSON. Using collections as well as folders.  Shows an example using flash to work with collections and some its ordering and tags. It also generates RSS feeds by default for all collections. Demos it in Facebook using VuVox and the RSS feed

So I asked for a compare and contrast xDrive toS3. I think the long and short of it is that S3 is geared towards app developers and the space belongs to the app – while xDrive belongs to the user and is a way to share data by the user. Has things like sharing, ACLs,  and other neat features. Then I asked them about bandwidth and they said it is free – I asked could I host my tiles for google maps or VE on it and they said – “If you think your bandwidth usage will be similar to a typical user” then you should be ok. There is definetly some fuzz here but the price is right, here are the terms of service (http://www.xdrive.com/terms.jsp) . I say Mr. Flood should move one of his tile sets to xDrive and we should see what happens.

Overall I was impressed since I didn’t even know that AOL had a dev network. There was a heavy emphasis on not being a walled garden – and I, being the eternal optomist, would like to believe. They definitely have their work cut out for them in convincing people but it could be really interesting. They need some better outreach to the dev community though I am not sure how. Perhaps I will play….

BungeeLabs in the afternoon

Had an excellent brownie at lunch – moist and chocolately…

BungeeLabs

They have a fusion server – demo – read an email on their fusion server – updates your outlook but also looks to salesforce to see if you have a lead or contact with that person. Doing SOAP calls behind the scene. Does address book against lookup against salesforce and exchange in real time. They are showing this as a reference app because they actually sell the server pieces that let you wire all this together. They have some HOT javascript going on between and roundtripping going on with their controls – add a address in one browser window and it updates automagically in another browser window. Very nice.

Another great demo with contacts and google maps – dragging and dropping contacts and getting mapping and routing.

The IDE to build apps is an Ajax application and is done only on the server. They have some great UI going – for developers it resembles VisualStudio or netbeans but it is a web app. You can bound controls to the same field as so they share state in real time. They have versioning built into the system and you can choose which to deploy from the source control. Pieces of the functionality can also be shared with other developers as controls with different licenses. You can also build a team and share the source and build.

The whole thing is running in BungeeLogic which is a C like language but written specifically for these on-demand type applications. Their IDE is hosted in bungee as well so eating their own dogfood.

They use a utility based model – they charge you for hosting the app based upon how many hits you get. No charge to develop.

If you want to add a new javascript api widget to your app you need to wait on them to write it. Said it is on their roadmap to allow developers to add them but not now. Can work with XML over HTTP or REST services without their intervention.

Really interesting tech and ideas – they definitely have some issues left to grapple with, such the behind the firewall deployment, but the stuff they have looks hot…

Day two of mashup camp

BungeeLabs was delayed due to traffic so this morning is an IBM fest. I saw their demos yesterday. They are making a big push to provides feeds (federation, repurposing, wiki-fying) in a corporate type setting. Not sure how this all plays with the other parts of IBM software stack so maybe we will learn more. Probably going to be the closed source version of SnapLogic or Kapow.

Dan Gisolfi is leading off the talk. Going to cover mashup ecosystem – IBM tech overview – QEDWiki code walk – Mashup Hub Code Walk. One of the points of this new “mashup” tech is to allow for more of a self service business pattern. Even though you may not be IT you still need to create a web site that combines information in new ways. [note - maybe this should really be called Portals 2.0] He says the data is the gold nugget – look at accuweather, dow jones… they are trying to find new channels for their content (i.e. data). How to drive revenue from this? His slides are wayyyyyyy to busy – great talking and ideas but his slides are distracting.

Interesting Widget ecosystem concepts chart – IBM is targeting the knowledge worker with widgets that even above standalone self-contained widgets such as Zude or Google gadgets – a widget of aware where it is on the page. He says they are not there yet.  Demo of a situation aware application for the EPA tracking Avian Influenza – wiki page with widgets on i. Not page aware widgets but from various places on the web – dapper widget, feeds…  Using a trulia widget with real estate and zip codes AND lat long (the key part)  along with accuweather wind patterns to look at bird migration patterns for spread of virus. Dynamically wired together on the page.

Next example mashing up trulia, dapper pulling stuff from greatschools.com,  and google maps. Very nice integration…

Another slide with a lot of the pieces of the ecosystem. Maybe there is a place where they could put these slides up so ppl can see them. My phone camera is not good enough to capture it so I will ask them to make it acccessible.

QEDWiki – mashup maker/hosting
MashupHub – catalog and collaboration for feeds and widgets with rating and tagging – embraces  OpenSeach (hooray for DeWitt)
DAMIO – browser based tech  that enable users to create custom feeds – like Pipes or SnapLogic
Ms. RITA – utilization management service to track widget, their use, and should we meter it for revenue (not always about money)
The platform is called info2.0

most of it is written in PHP with some having Java versions. The people deploying and using this stuff are really web developers and so PHP plays well in this space. It is again a split in big corporate (IT setting standards) versus smaller shops where PHP is much more common for rapid development. Another interesting note – if you look at most of the widgeters and content providers they have APIs in almost ALL the languages. Not dictating what language you use to consume. Some other large vendors should think about this as well….

Interesting note – yet again we are coming up against the pro versus amateur with tools discussion. Use case given is a knowledge worker who doesn’t want to wait for IT to write a custom report from the DB. So if the Knowledge worker has DB privs then they can use SQL to create a feed that can then be repurposed.  Predicition – get ready for the debate about who is expert enough to do this kind of stuff to grow here as well. Actually I am sure it already has and I probably missed it somewhere…

Other talks but I need to take a quick break…

More afternoon sessions

Intel and LignUP

Not sure that this one is going to be about – here we go. Community software managers help developers be successful on the intel platform. Talking about platform and context awareness. Intel has a web 2.0 tdk. C++ plugins for FFand IE http://intel.com/software/web20tdk

LignUp – App server enables developers to integrate real-time communication into web apps , portals, and process without CTI or SIP experience. Demo showing the server calling the devs phone when his laptop storage when it went below 16% free space. It called him and then connected him to the closes Best Buy so he could go buy an external hard drive. Just did another demo that when the server went above 60% CPU usage it initiated a conference call with an audience member. It is javascript and ASP.NET with VB.NET. They also have a facebook plugin that allows you to do text to speech or live record and then send to phones.
Microsoft – windows live services

Scott Mauvais – technical architect – kinda like tech marketing where they help prototype customers solutions. Going to show how to combine live properties instead of mashing other sources. It is going to be a mapping app – Look at photos for your and your contact and geolocate all of them.

Going to show single sign on, storage for photos, contacts, and a map. Man, doing all these live internet demos is scary – the MS guy is having problems but it works now. Showing adding pictures to a location. Both MS and AOL do prompting for permission to do stuff with the data.

Doing some code snippets to build some pages – all java script and simple html. Pretty slick – they have ACLs in their spaces (images or blog) that the javascript can take advangtage of privacy settings. Nice app he put together with fairly little work.

They are definitely going after ESRI with the internet mapping. He demos a oil fields visualization in the Gulf of Mexico that does a advanced visualization, asset managment, and workflow.

He then also states that he (and not the product team) thinks that the live properties are going the way of Office – not so great when it started but excellent integration among the different pieces. So given all the stuff that happens in their live apps they are going to leverage integration. They are also thinking of integration this with Office live server.

GOOG Mashup Editor

Trying to be able to create mashups quickly and easily. Need to handle feeds, give a UI, and give infrastructure.
All feeds are in  Atom like GData
UI Controls need to be cross browser compat
Simple declaration of UI components like HTML
Simple data modeled after a feed – so all data is from a feed
Provide a sandbox with one click deploy

They use <gm: tags to add modules and to create templates

Demo : they have an editor with specialized tags for google mashup things – it then gets compiled on the server and then hosted. Pretty easy syntax. This is how you also build gadgets.  Another demo showing how you can make an editable feed that you can edit your own data. Now a demo to create a list and a map – state parks listed with maps and google base for camping areas. Now google is having the live on the internet issues.

Missed most of jMaki and now Kapow

I am sad I missed jMaki - there seems to be a ton of great things happening there and it would be good to get a brain dump from one of the devs. I will will hunt Greg down today or tomorrow.

On to Kapow
Looks like they are going to be another company providing APIs to the stove pipes internal to enterprise organizations. They are going after the long tail of data and applications – small apps and data that not used by most people at the org but used consistently by people in the org and there are a lot of these little apps. They take structured and unstructured data and go back and forth to common mashup protocols. They also do web scraping. They have an open playing space and they also have Java and .NET APIs. Demo time – making an RSS feed by scraping an MLS site.

Next up is Zude
These guys allow you to create a portal where you mashup all sort of content from throughout the web. I am starting to fade in the day and so it is harder to me stay focused on the talk. He is talking mostly about why they did what they did and what the background is. I saw a demo of this during the speed lunch and it is pretty neat – he was taking all sorts of content in all sort of formats and adding it to a page as widgets. This goes beyond just adding RSS feeds. Here is some of their white papers and technology behind zude, fifth generation. They say it is all open source and so we should be able to use it and it supposed to be fast. Claiming that it is real time and small footprint.