Oracle Announces Roadmap for Plumtree / AquaLogic / WebCenter

UPDATE 2: I’ve incorporated all the great feedback and comments from ex-Plumtreevians, ex-BEA and ex- and current Oracle folks.

UPDATE: A bunch of Plumtreevians are contributing really good comments on this post over on Facebook.

bea_think_oracleI worked at Plumtree Software, Inc. from June 1998 to December, 9th 2002. In four-and-a-half years, the company grew from 25 employees to over 400 and it had thousands of happy customers before it was purchased by BEA Systems in 2005 for $220M. Here at bdg, we’ve been supporting dozens of Plumtree/AquaLogic Interaction (ALI)/WebCenter Interaction (WCI) customers since we opened our doors in December of 2002.

Back around 2005, BEA’s BID (Business Interaction Division) still had a lot of really smart engineers from Plumtree working on a lot of really interesting things, including Pages (think CMS 2.0), Pathways (kind of an enterprise version of and Ensemble (the portlet engine/gateway, minus the overhead and UI of the portal itself).

They were also working on an enterprise social network, kind of a Facebook for business if you will.

However, there was a lot of wrangling at BEA, primarily between BID/AquaLogic and BEA’s flagship product, WebLogic (the world-class application server). Most of the strife came in the form of WebLogic Portal vs. AquaLogic/Plumtree Portal nonsense. Senior management at BEA, in their infinite wisdom, had taken a “let’s try not to alienate any customers” policy and in the process they confused all their customers and alienated/frustrated quite a few of them as well. They renamed Plumtree to AquaLogic User Interaction (ALUI), put in place a “separate but equal” policy with WebLogic Portal (WLP) and spewed some nonsense about how WLP was for “transactional portal deployments” vs. ALI for .NET and non-transactional portals, but no one, including BEA management, had any idea WTF that meant. To further confuse the issue, the WLP team, which also had a lot of really smart engineers, built products like “Adrenaline” (which was basically a less-functional and more buggy version of Ensemble) rather than do the unthinkable and integrate Ensemble into WLP so that WLP could finally host non-Java/JSR-168 portlets.

I was really pissed about BEA’s spineless portal strategy, their “separate but equal” policy between WLP and BID/ALUI and their waste of precious engineering resources in an arms race between WLP and ALUI rather than just stepping back, growing a spine, and coming up with a portal strategy.

Because I can’t keep my pie hole shut, I started several loud, messy and public fights with BEA management. Why? Because the real loser here is the customer.

And BEA, because management got mired in politics and chose to waste engineers’ time on in-fighting and competition instead of building enterprise Facebook, which Steve Hamrick and I arguably already wrote in our spare time. All they needed to do was product-ize that and they would have owned that market.

In 2008, Oracle inherited this clusterfuck of a portal strategy when they bought BEA for $7B+, giving me new hope that cooler heads would prevail and fix this mess. The first thing they did was fire all the impotent BEA managers who were afraid to make any decisions. It took Oracle a while, but alas, they have finally arrived at a portal strategy that makes sense. I first learned about this strategy when I crashed the WebCenter Customer Advisory Board last Thursday.

First of all, let me say this: under the leadership of Vince Casarez, current (and future) customers are in good hands.

I realized when he said “everyone still calls it Plumtree” that this was going to be a bullshit-free presentation.

He also said something regarding the “portal stew” at Oracle that puts all of my ranting and raving in perspective: “Oracle did not buy BEA for Plumtree or WLP, just like it didn’t buy SUN for SUN’s portal product.” To rephrase that, Oracle bought BEA for WebLogic (the application server, not the portal) and Sun for their hardware (not for Java, NetBeans and all the rest of Sun’s baggage).

So, let’s face it, portals are a relatively insignificant part of Oracle.

However, they’ve finally did what I called for 2008 and what BEA never had the wits to do: pick a single portal strategy/stack and stick to it. SO, if you’re a current Plumtree/ALUI/WCI or a current WLP customer, you have a future with Oracle.

Here’s the plan, as I understand it.

All roads lead to Web Center (not Web Center Interaction, but Web Center)

At the heart of Web Center will be WebLogic’s app server and portal. Plumtree/ALUI as a code base will be supported, but eventually put into maintenance mode and retired. You get nine or twelve years of support and patches (blah blah blah) but if you want new features, you need to switch to the new Web Center, powered by WLP. CORRECTION: WebCenter will not be “powered by WLP.” At its core will be the Oracle-developed, ADF-based WebCenter Portal running on WebLogic Server.

All the “server products” (Collaboration, Studio, Analytics, Publisher) will be replaced by Web Center Services or Web Center Suite

Publisher will be subsumed by WCM/UCM (Web Content Management / Universal Content Management, formerly Stellent). The other products will be more-or-less covered by similar offerings in Suite or Services.

What about Pages, Ensemble and Pathways?

Pages is dead as WCM/UCM does it better. Pathways is getting rolled into the new Web Center somehow, but I’m not sure how yet. Perhaps I can follow up with another blog post on that. Ensemble has been renamed “Pagelet Producer” — more on that below. CORRECTION: Pathways is now called “Activity Graph” and it will be part of the new WebCenter. Think of an enterprise-class version of the Facebook News Feed crossed with Sales Force chatter and you’ll be on the right track.

What about .NET/SQL Server, IIS and everything else that isn’t Java?

This is a really interesting question and the key question that I think drove a lot of BEA’s failure to make any decision about portal strategy from 2005-2008. Plumtree had a lot of .NET customers and some of the biggest remaining Plumtree/ALUI customers are still running on an all-Microsoft stack. In fact, one of them told me recently that they have half a million named user accounts, two million documents and 72 Windows NT Servers to power their portal deployment.

So, let’s start with the bad news: Oracle doesn’t want you to run .NET/Windows and they REALLY don’t want you to run on SQL Server.

(That will change when Oracle acquires Microsoft, but that’s not gonna happen, at least not any time soon.) WebLogic app server and WLP/WCI, to the best of my knowledge, will not run on SQL Server. They will, however, run on Windows, but I would not recommend that approach.

It’s inevitable that large enterprises will have both .NET and Java systems along with a smattering of other platforms.

So, if you’re a .NET-heavy shop, you’ll need to bite the bullet and have at least one server running JRockit or Sun’s JVM, one of Oracle’s DB’s (Oracle proper or MySQL), WLS/WLP/WCI and preferably Oracle Enterprise Linux, Solaris or some other other flavor of Un*x. CORRECTION: WLP will run on SQL Server. Not sure about the new WebCenter Portal, but my guess is that it does not.

Now, for the good news: the new WCI, powered by WLP and in conjunction with the Pagelet Producer (formerly Ensemble) and the WSRP Producer (formerly the .NET Application Accelerator) will run any and all of your existing portlets, regardless of language or platform.

This was arguably the best feature in Plumtree and it will live on at Oracle.

.NET/WRSP and even MOSS (Sharepoint) Web Parts will run in WebCenter through the WSRP Producer. The Pagelet Producer will run portlets written in ANY language through what is essentially a next generation, backwards-compatible CSP (Content Server Protocol, the superset of HTTP that allows you to get/set preferences, etc. in Plumtree portlets). So, in theory, if you’re still writing your portlets in ASP 1.0 using CSP 1.0 and GSServices.dll, they will run in the new Web Center via the Pagelet Producer. Time for us to update the PHP and Ruby/Rails IDKs? Indeed it is. Let me know if you need that sooner rather than later.

How do I upgrade to the new WebCenter?

Well, first off, you have to wait for it to come out later this fall. Then, you have to start planning for what’s less of an upgrade and more of a migration. Oracle, between engineering and PSO, has promised to provide migration for all the portal metadata (users, communities, pages, portlets, security, etc.) from Plumtree/ALUI/WCI to the new Web Center, with WLP at its heart. (Wouldn’t it have made sense for some of those WLP engineers to start building that migration script in 2005 instead of trying to compete with ALUI by building Adrenaline? Absolutely.) All your Java portlets, if you’re using JSR-168 or JSR-286, will run natively in WLP through a wrapper in WebCenter Portal. Everything else will either run in the WRSP Producer (if it’s .NET) or in the Pagelet Producer (if it’s anything else). The only thing I don’t fully understand yet is how to migrate from Publisher to UCM, but I’m due to speak with Oracle’s PSO about that soon. Please contact me directly if you need to do a migration from Publisher to WCM/UCM that’s too big to do by hand.

The only other unanswered question in my mind is how the new WebCenter will handle AWS/PWS services — the integrations that bring LDAP/AD users and profile information/metadata into Plumtree/ALUI/WCI. I wrote a lot of that code for Plumtree anyway, so if Oracle’s not working on a solution for the new Web Center, perhaps I can help you with that somehow as well. CORRECTION: User and group objects are fully externalized in Web Center, so there is no need for AWS/PWS synchronization. (Thanks, Vince, for pointing that out.)

So, that’s my understanding of the new portal strategy at Oracle.

Kudos to Oracle’s management for listening to their customers, making some really hard decisions and picking a path that I think is smart and achievable.

I’m here to help if you have questions or need help with your portal strategy or technical implementation/migration.


(Some other notes about discussions that have spawned from this original post.)

Q: What’s the future of the Microsoft Exchange portlets (Mail, Calendar and Contacts) and the CWS for crawling Exchange public folders. Retired and replaced with something Beehive related? Still supported? For how long? Against what versions of Exchange?

A: We’ve got updated portlets for Mail & Calendar in WebCenter now for Exchange 2003 & 2007. We don’t have a Contacts portlet but it could be added quickly if we see a large demand. Crawling public folders can be done with an adapter we have for SES [Oracle Secure Enterprise Search] already. We’re working but aren’t done with a new version of KD on top of the new infrastructure that will come out post PS3. (Contributed by Vince Casarez.)

Q: If migration scripts are provided to move WCI metadata into WebCenter, I understand that a portlet is a portlet, but what about pages and communities, users and groups, content sources and crawlers, etc.? Do they all have analogous objects in WebCenter or is there some reasonable mapping to some other objects?

A: Pages and Communities follow a model where we extract/export the meta data and data, then run it through a set of scripts that create a WebCenter Space for each Collab project/community and a JSPx page for every page. Users and Groups will come out of the LDAP/AD directory they are already using and the scripts associate the right permissions to each of the migrated objects. I don’t recall what we did about crawlers but since we use SES directly, all the hundred or more connectors we ship for SES are now available for direct usage. The scripts go through a multiphase approach to move content, then portlets, then pages, then communities so that dependencies can be fixed up versus trying to do a manual fix up. (Contributed by Vince Casarez.)

Q: Will any existing WCI-related products that are slated for retirement (e.g. Publisher, Collab, Studio, Analytics, etc.) be re-released with support for Windows Vista, Windows 7, IE 8, IE 9 or Chrome?

A: For Publisher, we are planning a set of migrations to quickly move them to UCM. For Collab & Studio, we have new capabilities in WebCenter Spaces to match these functions. For Analytics, we’ve also rebuilt it on top of the WebCenter stack with over 50 portlets for the different metrics and made sure we provide apis/ access to the data directly. These analytics data also feeds the activity graph in providing recommendations for people on the content and UIs that are relevant to them. These are tied into the personalization engine that we brought over from the WLP side. So there is a rich blending of the best features from WLP with WCI key features. As for Neo [the codename for the next release of WCI], we are certifying the additional platforms. On the IE 8 front, we’ve just released patches for WCI 10gR3 customers to be able to use IE8 without upgrading to Neo. (Contributed by Vince Casarez.)

Lifehack: Free or Cheap SaaS Tools I Used to Get to Inbox Zero

This article original appeared as a guest post on Scott Abel’s blog, The Content Wrangler.
Lately I’ve been really overwhelmed by my email inbox. This is not a new problem, but in the past I’ve been able to keep it at under a hundred emails; recently it has grown to nearly 300 and it has really begun to interfere with my getting things done.

So, last night, I took a good, hard look at what was really IN my inbox.

About 40% of the notes consisted of links sent to me by well-meaning people who thought I should check them out for various reasons. Another 30% were suggestions on how to make our products, marketing materials, services, etc. better from employees, customers, partners and other well-meaning people. Of the remaining 30%, about half were personal introductions to potential partners, customers, investors or other people with whom the authors thought I would want to connect. The other half were ‘to-do’ items of a business or personal nature, some sent by me to myself (ick!) or by other people.

I think maybe one or two messages actually consisted of correspondence — by that I mean something like the letters of yesteryear that we used to send through snail mail. It’s interesting to see how the bastardized email of today is so different from the purpose for which it was invented, but that’s the subject of a whole other article. However, while I’m digressing, it’s worth noting that

email functions brilliantly as a “better matchbox” than snail mail, but at the same time it performs really poorly at all the other functions that it’s used for today.

Email is not a contact management system, a customer relationship management (CRM) system, a link-sharing/social bookmarking tool, nor a support ticketing/issue tracking system. Not by a long shot.

The goal for me was to put all these messages that shouldn’t remain as emails into their proper home so I could deal with them appropriately while maintaining my sanity.

Now that I had performed some analytics, it was time to get organized! Here are the tools I used to clean up the mess: Basecamp, Highrise and Instapaper. Instapaper is free; however the 37signals products Basecamp and Highrise carry a small monthly fee.
[Note: They also have trial versions, but don’t expect to get too far with them since 37signals made the free versions just useful enough to show you their value without actually providing any.]
Getting from almost 300 emails to under 20 took about two hours and it was time well spent. I made one pass through my bloated inbox and took one of these actions, based on the type of email:

Email Type #1: “Hey, you should check out this link because. . . .”

Opened the link and used the “Read Later” bookmarklet from Instapaper to save the link for when I have to time to read it. If the email containing the link had something interesting in it (besides the link), I copied that into the notes field for that link once I had saved it to Instapaper. If you care to share what you’re reading/bookmarking, you can also use a bookmarklet for this. I find Instapaper easier though, because you can bookmark a link with one click. forces you to enter tags and other metadata, which increases friction and slows down the process of bookmarking.

Bottom line: Bookmarking, per se, is a simple, rote task that shouldn’t take more than one click to accomplish.

Email Type #2: “Hey, you should make your product better by doing this. . . .”

Read the email. If there were specific action items associated with it, I created to-dos in Basecamp (under the project for the appropriate product) so that we can address them in a future release. We maintain a to-do list for each release of each product and another to-do list that serves as a backlog for each product. (Some agile tools refer to this as “the icebox.”) When we’re planning a release, we pop the most important things out of the backlog and move them into the current release to-do list.
If the to-dos were general, more thematic suggestions without specific action items associated with them, I copied the suggestions to one of our design writeboards in Basecamp. Then I responded to the email thanking them for the feedback and deleted it.

Bottom line: Product feedback and support tickets belong in Basecamp or a support ticketing system … or even a CRM, but they should never be kept in email as email is not the right tool for tracking the support ticket cycle.

Email Type #3: “Hey, you should sell to (or partner with) so-and-so. . . ”

Forward the email to Highrise’s email dropbox. Delete. Done. When I process my Highrise queue of messages, I can decide whether or not to pursue these leads on a case-by-case basis. Sales leads belong in your CRM system so that they can be tracked and managed. Email is the wrong tool for tracking the sales cycle. If you want to close sales deals and you’re using email as your CRM system, important communiqués are going to slip through the cracks and you’re going to lose business as a result.

Bottom line: E = mc2 but Email != CRM.
Email Type #4: “Hey, Chris, meet so-and-so. Hey, so-and-so, meet Chris”

Reply All and start the process of scheduling a good time to talk. However, there’s a bit of a hole in this, because if I then delete the message, how do I ensure that so-and-so and I actually end up talking/meeting? If you have any suggestions about how you’ve solved this problem and what tools you’ve used (besides stinkin’ email), please let me know in the comments field associated with this blog post. I guess I could use our CRM for this, but that’s kind of like using a bazooka to kill flies.

Bottom line: I don’t know what the best tool for this is, but I do know that it’s most definitely not email.

Email Type #5: To-do item (not related to a product or a lead)

Put in on my to-do list. Right now, somewhat ironically, this is an email that I keep perpetually in draft status. To-do lists are a funny thing. I’ve used Remember the Milk, Google Spreadsheets/Documents and a number of other tools, but frankly, nothing beats a text file. By keeping it as a draft email in Gmail, I always have access to it from anywhere, buy you can easily accomplish this with Google Docs too, or a number of other tools.

Bottom line: Your inbox should not be your to-do list. Use a text document, a to-do management tool or even a piece of paper and a pen. There’s something inherently gratifying about the physical, visceral action of scratching something off my to-do list with a big, fat marker (preferably a Sharpie). No tool I have encountered can come close to emulating that feeling of accomplishment.

Email Type #6: Personal Correspondence

Print it on nice paper, frame it and hang it on the wall! Seriously, these have gotten so rare, that I really don’t mind them at all.

Bottom line: This is what email was designed to do, so feel free to use it for that. Enjoy it, because your friends would probably rather update their Facebook status than send you an email. If they do send you emails (and there’s no to-do/action-item associated with them), then they’re a true friend. You should return the favor with a personal email of your own, or, if you really want to surprise them, drop a handwritten note to them in the postal mail, preferably with a designer stamp that reflects your sense of style.

There’s something really sexy about being retrosexual — try it, I guarantee you’ll get great results!

Conclusion: I didn’t quite reach Inbox Zero before my head hit the keyboard, but I am down to under 20 emails in my Inbox. Every time I hit “delete” I could feel my stress level, my blood pressure and my state of disorganization decreasing proportionately.

So, how many messages are in your inbox? What do you think of my approach? What tools and strategies do you use to manage all this email insanity? I’d love to hear your comments. Just don’t email them to me! :-)

Not resting on our laurels

Feedhaus_StickerTeam Feedhaus put out a new release tonight. This was primarily a bug-fix release. Here’s a run-down:

1) Speed up history detail pages — we’ve got more than 1.2 million stories stored up since our private alpha began in early September. Finding those records to show required some heavy lifting, but by adding some indexes to that table and making a few code changes, you can now play with the history slider all you want and pull up old stories almost instantaneously.

2) Speed up regular detail pages — some browsers were having trouble displaying some of the detail pages, mostly just because they were too big! We’ve capped the number of stories now at 20 per detail page. Eventually we’ll add pagination, but for now, there’s still plenty of news and blogs for you to read, with the latest and greatest stuff always bubbling to the top.

3) Detail page clipping — some tags returned feed content that was too wide for the detail page, which required you to scroll horizontally to see the full extent of the content — yuck! (We’ve had lots of complaints about this.) Using some CSS hacks, we fixed that for all the major offenders that we could find. If you find any other “wide loads” (i.e. detail pages that are too big for their britches), please let us know.

Feedhaus Selected as a SXSW Web Award Finalist

2008_web_awardsWe’re very pleased to announce that Feedhaus has been selected as a finalist for a prestigious SXSW (South by Southwest) 2008 Web Award!

We’ve been selected, along with four other great nominees, from among a pool of hundreds of sites for the “Technical Achievement” category, which, according to SXSW, describes “sites that are re-inventing and re-defining the technical parameters of our online experience.”

As a finalist, we’re also eligible for the “People’s Choice” award, so please vote for Feedhaus (daily until March 3rd).

Kudos to the SXWS committee for recognizing Feedhaus! Hope to see you in Austin from March 7th-11th for this great conference.