MacPorts and Mavericks

Upgrading to a new Mac OS X version always brings compatibility issues with Mac Ports and Mavericks is no different.

I am currently fighting through my version of Dante's Inferno trying to upgrade all my ports.

I started with good intentions and did all the corrects steps for an upgrade:

1. I updated the MacPorts Installation itself.
2. I updated Xcode and the command line tools (very important as the TCL support has changed).

Then the fun started.  Everything broke on the llvm-3.0 installation, which is not supported in Mavericks.  So far I have been backing out ports that depend on it until I ran into python27.

I deactivated llvm-3.0 and all its dependents and then did an

"upgrade python27"

That totally blew up on the compiling the Fortran option for Atlas.  That's where I have stopped for now.

I think I found a path!  I did an "upgrade gcc45" (for the Fortran) and now everything is flowing.

Update (11/3/2013)

I was able to get the general ports base updated/upgraded, but had problems with asymptote and octave-devel.  I opened tickets on both.  I worked with the octave maintainer this week and it now compiles successfully.  According to him the octave base is compiling with a Fortran compiler (for now?).  Asymptote is still broken.

Modern Web Development

It should be no surprise to most, but web development has been completely transformed in the last couple of years.  The biggest change from my perspective has been the rise of Javascript-enabled front ends.  An even bigger change is non-web networked apps in the form of smart-phone apps, but that is for another day/post.

I have been journaling Javascript libraries:

Here is my starting point:

• node.js
• jquery
• backbone.js
• jgraph
• jointjs
• draw2d
• GoJS
• yworks
• raphael.js
• wireit
• yahoo pipes
• jit
• jsPlumb

I am sure there are more.

Working with RabbitMQ and MacPorts

I have installed RabbitMQ with MacPorts on my Snow Leopard laptop.

I have run into a problem with the server restart after a reboot.

The launchctl agent provided by macports starts the "rabbitmq@localhost" but the node created during the installation (with the erlang cookie) is set to the "rabbitmg@machinename".

I think that there are two solutions: one, alter the launchctl agent (unloading it, changing it and reloading) or, two, setting a rabbitmq-env.conf file and forcing the node name to be equal to the installed node. I chose the latter and it worked.

Installing Django on Mac OS X Leopard

I am building a Django application that has to be hosted on a Leopard Mac OS X Server.

The road to success has not been easy. I first went down the path of using MacPorts and that lead to difficulty. I ran into well-know problems of 32 bit versus 64 bit conflicts between MySQL, MySQLdb and Python.

To date I have found an equilibrium by installing MacPython 2.6.6 and the 32bit version of 5.1 MySQL. So far so good.

I have installed mod_wsgi and will attempt next to publish the Django application next.

Email in the SMB Segment

The Future of Email in the Small and Medium Business Segment

Introduction

Email was the original internet “killer application” and remains as one of the most important business applications in the modern office. The ubiquitous use of email has lead to the situation where email is considered both a commodity and a utility. In this role, email suffers from considerable cost pressure as a commodity and serves as a platform for many complex business functions, such as customer relationship management (CRM).

Two independent forces are forcing all businesses regardless of size to reconsider their current operational footprint for email. The first is compliance with Federal Rules of Civil Procedure (FRCP) or more commonly known as eDiscovery. These new rules contain specific requirements regarding document retention, including email messages, for use in the civil litigation. The second force is the industry-led consolidation of communication technologies whereby telephone, instant messaging, wireless phones and email are converging to common platforms.

In this research paper, we analyze the changes occurring in the small and medium business segment and make recommendations that maximize business value for the small and medium business owner.

How is email changing?

Email service, one of the original internet applications, is an indispensable part of any modern office. From a technology point of view it is relatively mature. Over the last few years issues in email have focused primarily on cost of delivery and email security, either in the form of spam or viruses. Other issues such as mailbox size have been somewhat insignificant as email has benefited from the Moore’s Law effects associated with the decreasing costs of disks and storage. Email providers have focused primarily on cost effectiveness, security services and operational effectiveness as market differentiators.

Two new market forces are now changing the use and cost of email. The first comes under the header of messaging convergence, which is getting all your messages “anywhere, anytime.” The second big market change is the expanded use of email archiving to meet federal regulations for document discovery in civil litigation.

Messaging Convergence

In the enterprise market segment, you cannot turn around without seeing titans like Microsoft, Cisco or Google touting Unified Messaging or Unified Communications. The vendors want to merge voice, email, instant messaging and cellular communications into a single platform. The theory is that the cost of synergy outweighs the cost of change. The battle in this arena is early and it is still very unclear who will be the dominant player and, more importantly for this white paper, how email will be impacted.

eDiscovery and Archiving

Archiving has the most straightforward affect on email. To respond to discovery requirements, there are three primary ways to comply with document discovery:

1. Rigid document retention rules, such as delete after six months
2. Save everything and buy storage or archiving systems to accommodate the data
3. Employ selective archiving with an archiving system

There are pros and cons to each approach, but most people prefer the second option because it is the most user friendly.

In the enterprise market space, the “infinite” archive as a document repository is mostly a factor of money and scale. In the small-to-medium segment, many companies operate in a POP3 environment which makes archiving impossible. In response to this business condition, many email vendors are now providing archiving service as added option for their POP clients.

Considerations for Change

Do you have to make a change in your email system? The answer is mostly likely “yes” even if you are happy with your service, especially the price. The simple answer is the FRCP regulations cannot be ignored. What other factors are important if you have to make a change? In this section we will deal with important and perhaps urgent issues that you should consider as you evaluate changes to your email service.

Archiving

As stated above, new FRCP regulations necessitate your consideration of how to handle document retention. The first, and most important, step is a call to your corporate attorney with a frank discussion of what risks you practically have for litigation. Drawing the line of how many documents (emails) to save and archive is highly subjective and somewhat a matter of philosophy. The ever decreasing costs of storage have typically lured people into saving documents rather than destroying.

Large corporations with permanent legal staff also push for specific archiving systems that aid in the preservation of documents identified during the course of litigation. This topic should be discussed thoroughly with your attorney. The simplest approach is to save everything with a generous destruction window, such as seven years.

POP3

Many companies in the SMB segment use a POP service because it is the simplest and cheapest way for a hosting provider to email connectivity. Using an IMAP service, where the email stays on the server, has typically been reserved for larger corporations.

The decision to abandon POP versus a server-based email system is largely interconnected to two issues: what type of email client (like Outlook) do you want to use and what is your archiving strategy?

The cost of Hosted Exchange, for example, is not getting extremely cheap and a small company can enjoy all the features of a server-based email system for the same costs of a POP server a few years ago.

Calendars

If you use group calendaring, then odds are likely you are already using a server-based email system like Exchange. If you want to use group calendaring, you need a server-based email system (and thus will abandon POP).

Smart Phones

Smart Phones are providing large productivity gains in the modern office and are becoming near ubiquitous. Blackberries and iPhones can be integrated deeply into the email system.

Convergence

Unified Communications is definitely in the future for all email users. Aggressive, early adopters in the SMB space may consider a UC platform, but waiting for some maturity and vendor shake-out is the best option for now.

Options and Recommendations

The recommendations below assume an office size of 20 people and that smart phones are essential. Other factors will be considered below.

All recommended designs assume a hosted vendor solution. Except under extraordinary circumstances, groupware-style email installations under 1,000 users are better suited for a hosted solution rather than an internally provided solution.

Technical Design

The most important factors in selecting an email service are the use of group calendaring and “tight” smart phone integration. If your organization uses group calendars, groupware mail services, like Lotus Domino or Exchange are required. If you want your smart phone to synchronize all your email, folders and contacts, you also will require a groupware email service. Almost all email services will synchronize incoming email with a smartphone but outgoing email and contacts either cannot be done or require a third party application. Although native email client synchronization may not be possible with non-groupware services, most smart phones have web browsers that provide real-time access to the user’s mailbox.

The next key element of the technical design is the selection of an archiving option. In this area, the market has clearly spoken. Most vendors are offering two types of archiving. The first is the “unlimited” or extremely large mailbox option. The second is an unlimited archiving service, which normally comes with simple litigation tools like search and tagging.

A final consideration in a technical design is a company’s commitment to or comfort with a particular vendor’s products. For example, if a company is comfortable with Microsoft Office and Outlook, one will probably lean towards using an Exchange-based service.

Vendor Recommendations

The cost of non-converged email service is very cost-effective, especially if you want a non-groupware email service. For example, Rackspace Hosting provides fully hosted Exchange Email service with Blackberry integration for $22.50 per user per month, but offers IMAP non-Exchange hosted email for $1 per user per month and that includes Blackberry integration!

Groupware solutions

In this section, I will assume that the hosted groupware solution is Exchange. For companies needing groupware with smartphone integration, the recommended solutions are as follows:

Vendor	Email	Smart Phone	Archiving	Specifications
Rackspace Hosting	$12.50	$10	$3	2 GB mailbox
Microsoft	$14.00	$10	$4.50	5 GB mailbox
Intermedia	$11.95	$10	$49 for first 2 GB $5/GB thereafter	4GB mailbox
SherWeb	$8.95	$9.95	$45 for first 3 GB $5/GB thereafter	3 GB mailbox

Non-groupware solutions

If your company is comfortable with the restrictions of a non-groupware solution, there are some very compelling offerings. There are as follows:

Vendor	Email	Smart Phone	Archiving	Specifications
Rackspace Hosting	$1.00	$0	$3	10 GB mailbox
Smarsh	$12.00	n/a	$0	unlimited
Google	$50/year	n/a	$45/year	25 GB mailbox

Archiving solutions

If your company only needs archiving, several vendors provide a standalone service.

Vendor	Archiving
Google	$45/user/year
Smarsh	$48/user/year
Microsoft	$54/user/year

If a company has its own POP or IMAP email server, there are appliances available that perform archiving. One such example is the Barracuda archiving appliance. The smallest appliance holds 250 GB of email for approximately $2,000. Over a three-year period and an office of 20 people, the cost averages $2.77 per user per month.

Final Thoughts

The current direction for the SMB definitely should not include Unified Messaging except for the early adopters. When should most companies consider UM and are there other factors that should be included in an email change process. As for UM by 2011 or 2012, wider spread integration will emerge that unify these technologies for the SMB market.

I still believe that totally online office application suites are also in the early adopter phase. One other major area, however, that SMB companies should evaluate when switching email is shared storage or office tools like SharePoint. SharePoint, in fact, is quickly becoming a leading seller for Microsoft and may eventually become an equal partner to its Office suite of products.

Microsoft fires back!

Microsoft fired back at Google today in what is sure to the start of The Great Geek War of the 21st century.

Today Microsoft offered details for Office Web. Office Web will feature lightweight versions of Word, Excel, PowerPoint and OneNote. Office Web will be delivered in three modes: FREE to Office Live users (400 Million strong) on the Internet, via a subscription model for Office, or part of their enterprise, on-premise offering.

Office Web is a shot across the bow of GoogleApps. The online offerings of Google (and others) pale in comparison to the feature-rich local applications. Obviously, both ends of the spectrum have the points. Local applications tend to become bloatware. Online offerings are limited in user interface features. The Great Geek War is a race to see who can create the best balancing act between features and ubiquity of access.

In another post, I will discuss the real tripping point in this battle for user eye balls: data security.

Google OS: should we care?

This week was the official announcement that the Google OS was real and not just a figment of geekdom's imagination.

If you google "Google OS", there is literally an avalanche of punditry on all sides of this issue.

Should we care? Or, most poignantly, should Apple or Microsoft care?

The short answer is that everyone will care about the new Google OS effort. Here are some thoughts about how it might effect each group.

Apple

Traditional market and game theory would suggest that Apple has the most to lose in a Google OS beachhead. I disagree. Apple is moving towards network-based computing from the device end. This vertically integrated strategy has considerable strength because Apple can fine-tune the user experience based upon overall quality of networks and device computing power.

People have been proposing and postulating "net computers" for 15 years or more. No one has done it successfully because the user experience is just too good with local computing.

As the 3G networks become ubiquitous net computers will become possible and even desirable, but I think Apple can play a strong hand by building on top of its strong devices.

Microsoft

No monolith ever falls from top over night. This will be true for MS as well. Google OS is targeted directly on Microsoft.

I think the real reason that Google has countered with Google OS is that they perceive that Microsoft's Bing finally means that Redmond is serious (or competent) about Search. Google is an advertising company, first and foremost, and Google OS means that the platform vendor cannot shut out its advertising revenue.

I would look for search/advertising/operating system to merge for both these companies on the consumer side.

Will Google take significant share from Microsoft? Not initially, but look for Google to get at least 10% of the netbook sales in 2011.

I think that Google will have a harder time in the enterprise market. Nothing I have seen here looks like it solves business problems.

Consumers

Consumers get more choices. This is almost always a good thing. In the end portability of data will be the issue that people will be most worried about with any of the net vendors. How easy will it be to move my pictures from Flickr to Picassa? Can I move my Google Docs to Microsoft LiveOffice and vice versa? This FUD will slow adoption at first, but I suspect their will be a market for moving products that facilitate this lacking any industry standards.