Dog-Fooding

I was reading an article about Visual Studio 2010, next generation Microsoft developement toolset. What cought my eye was not the features. It was dog-fooding practice so wide spread in Microsoft. Dog-fooding or “eating your own dog-food”, is the practice of a company using its own products internally. Microsoft employees browsing internet on IE is doog fooding. It is trivial now that IE is a market leader, but it was true when IE was in its infancy and Netscape Navigator ruled the web.

It touched the nerve with me. I am dog-fooding. I use gganttic, my project mangement software, every day for my own projects.
Since gGanttic is based on Google office applications I made decision to use Google Documents and spreadsheets for everything. So I am eating Google’s dog-food full time.

What is obvious to me that Google Document developers are not dog-fooding.
I am dead sure the specificaiton of the next version of Google Documents is not written in Google Documents. The sheer amount of problems I face when writing a lengthy document with a number of images shows me that I am probably in a tiny nimority of Google serious users. And that Google developers are not in this minority, at least not many of them. For which I am sorry.

For their own sake, as for their users, they should start working with those applications for real.

Economic Crisys and Google Office

Economic crisys is upon us. All of us. Wherever we are. In the intangled world of today noone is exampt from feeling it. To cope one has to cut spending to what is necessary. This much is trivial.

I predict that in that environment more people then usual will want to try Google Docs which I like to call Google Office. Google got gMail and Calendar, Documents (word editor), Spreadsheets, and Presentation. They all are web 2.0 applications, which I call HTML/CSS/Javascript applications doing enough on client and use Ajax well enough to create good responsiveness.

So far I did not see users giving up on their Microsoft Office or even Open Office and marching on to Google Office.

I am working on a Google Docs integrated Project like appliction to fill the void Google left in that area. Our team made decision to use Google Docs exclusively. It was and still is a right decision. Because now I know first hands the good the bad and the ugly sides of them.

I belive that in down economy one should try it. It is free. I encourage you to do so. It allows collaboration, which is important when travel budgets are slashed.
On the same breath, I want to manage your expectations a bit. You might discover it is too early for you.

(Added on 2008-11-15, I just found the fresh Business Week article saying the same thing:

CHEAP TECH FOR HARD TIMES
By Steve Hamm

As the U.S. enters what appears likely to be a painful recession, a major shift is taking place in how businesses assess technology products. They’re under terrific pressure to cut costs. According to a newly revised forecast from market researcher IDC, growth in U.S. tech spending will decline to 0.9% in 2009, down from a previous forecast of 4.9% growth. But rather than just slice budgets across the board, many companies are switching to a handful of new technologies that save them money. These technologies existed during the last recession, but they were immature. Now they’re established, and the downturn seems likely to hasten their adoption. Chief among them are software delivered over the Internet, known as cloud computing, such as Google Apps…)

Google Office – the Good The Bad and The Ugly.

The good – it is online. It runs on every major browser and every OS. It passes the minimal requirements test in terms of usability and features. It is free. It is highly collaborative. It is, to a degree, integrated. It is constantly evolving and improving. It is versioned (Docs, Spreadsheets) meaning you can get back to old versions. It got API, meaning 3rd party applications can be added to make use of the core set and complement the missing features.

The bad – some applications are more mature then others. gMail, Calendar and Spreadsheets are better then the rest (In My Humble Opinion). Ajax, the technology Google use limits what is easy or even possible to do. No use of Flash or Silverlight makes it run on every browser with nothing else installed, but limits in many ways.

The ugly – sometimes the limitations of Ajax make you cry. Many bugs. Long documents are very hard to edit in Documents. For some reason Documents renamed my internal bookmarks. It all means that Google Office is for early adopters, not mainstream and not for enterprise where reliability is crucial.

I am prepared to live with bad and the ugly. I am set to improve it where I can (see www.gganttic.com). That is because it evolves. I see improvements all the time. Being early adopter I am more tolerant to rough edges, hoping to win by being first to ride the paradigm shift.

Wheter Google Office is right for you I can not say. And you can not say either, unless you try. Do it.

Good .NET hosting – DiscountASP.net

It is hard to find a good hosting service for your .NET application. I have had serveral, and kept looking until I found the one which I want to recommend.
It is DiscountASP.net (sometimes abbreviated as DASP by the community)

I used it for my personal domain, for couple of clients and now for my new startup. All is shared hosted.
I like the price, and because of it can live without phone support. I would rather them lower prices further then provide phone support, really.

I have opend couple of tickets over time, and was generally satisfied. Was long ago, so sorry I don’t recall specifics. I use FTP all the time and it is fast enough. I do stuff in control panel, define web applications etc, and it is responsive and logicaly organized.
One thing that I value is quick adaptation of new Microsoft technologies. I am on the bleading edge and need the latest stuff.
Another strength is site layout. I tried one even cheaper hosting at a company where I regsitered my domain and was very disappointed by their web site organization. I could hardly find the way to control panel, through pile of advertising and different other services. DASP focuses on hosting and is modest in their attempts to cross and up sell.

There were several small outages that I recall. With Amazon down twice this year what you can expect? Nobody is 99.999, i am convinced. DASP was IMHO well within the reasonably dependent service definition. Suffice to say my application breaks so much more then the hoster, that it will take me time when DASP becomes the reliability bottleneck.

Google Documents allow CSS editing

I have noticed that recently, Google Documents allow CSS editing.
It was there for a while, I simply did not do much of Google Document writing lately.
Now I am back into it. I am writing a long Tutorial document. This Tutorial ought to look well. Hence my interest in the subject.

People reaction to it, from the comment I have seen, range from positive – it allows customization of the look – to negative. The negative is interesting. Commenters point out that CSS is too technical for a general purpose Word processor. They claim GD is on the path to becoming an HTML editor. This rubs many technical readers the wrong way – to them nothing is wrong with HTML.

I personally adore CSS and what it allows me to do. One thing, before I start tweaking my TUtorial CSS, I’ll go out Googling for CSS libraries for Google Documents. I am sure someone did some good work already that I can benefit from.

There is, however, a point in the “too technical” argument.

Microsoft Word allows control over styles, but only via a hierarchy of dialogs. I would agree that CSS can be intimidating for newbies. That is why Dreamweaver, and it hard to get more professional in HTML then that, have dialogs to create/edit styles.

I assume there is somebody at google who monitors user reaction to new features.
If those who claim GD is “too technical” are the majority, then at some point we’ll see Style Dialog coming to fore, and CSS hidden somewhere.

Still being able to customize styles, one way or another, is, IMHO, a huge step forward.

Google Chrome – why it can’t be my default browser

Google is out with the new browser Google Chrome. The most important features, for me are – it has minimalistic interface and an run web applications in a window with no “chrome” at all.

Chrome is the lingo for the UI of the browser itself. IE and Firefox are crowded with menu items and toolbars. Each add-in you add, and some are very useful, so it is not dissing them in any way, crowds the chrome. Google Chrome kills all of the browser UI for web applications. I am writing this blog post in a window where only toolbar belongs to the browser. The rest is the webtop. So Google Chrome’s name is kind of tongue in the cheek thing. Google Chrome is chrome-less.

Another good feature is speed, with Google V8 JavaScript engine is said to work very fast, faster then FF3.

What disappoints me? Two things. One is my Silverlight 2.0 applciation will not work in the current implementation of Chrome. And the other is my Roboform will not work with it. Roboform is the password manager. I routinely sign into 30 web sites/ web applications. I use Google’s own Documents trying to move to Google “office” exclusively and do without Microsoft Office. To be able to sign in instantly is the key for that. Roboform allowed that. And Roboform does not work with Chrome yet.

There is an open source and free alternative to Roboform, called PassKey. I did not see reports it works with Chrome. If it does, I’ll switch over from Roboform to PassKey instantly.

I could have asked Chrome to remember my passwords. I don’t like it. I like to be in control, that is one. I want to carry my passwords around with me on the flashdrive that is another reason.
So, currently, I shall use Chrome to browse the net, and IE / FF to actually have my work done.

Python/Django or Ruby On Rails?

I am architecting a big multimillion user web service, version 2. It is going to be a complete rewrite of version 1. Its current implementation is PHP. I consider if to stay with PHP or move over to Python or Ruby. The service has reporting and administration modules so GUI framework is relevant somewhat.

My current choice is Python. And the reason why I choose it over Ruby is that I feel it is going to be much harder to find Ruby people then Python people.
I visited Israeli Ruby on Rails user group meetings. The people there are bright and passionate, but they are too few to count.

By the way here is how NOT to try and hire good Ruby developer.
Just funded company Confidela posts this killer line in job description for Server Architect
Advantage:
- Experience in Rubby On Rails

One should be really desperate to want to work for a company where people don’t know any beter or where hiring manager does not proof what HR people put on their web site.

A man in the Arena

Techcrunch is a place I (MK, gGanttic’s interim CEO) visit regularly. I am also subscribed to its daily email. In my value scale, that is the highest mark a web site ever gets.

I recalled his post “A man in the Arena” recently. You should read it, if you want to understand what drives us the team behind gGanttic. We are going to change the world or die trying.

Analogy of a tech entrepreneur to a gladiator is a good one, to me. Following it a bit farther, a gladiator needs armor and it is the money investors provide. gGanttic is pre-seed, meaning we are a Man in the Arena going against the horrors of the world armed with nothing more then our skill. How’s that for chutzpahs?

One might assume, judging on how crazy that sounds, we’re a typical teenage hackers with lots of time and a readily available basement. And you’d be wrong. We are all industry veterans with 15 years of experience each, held senior positions in software development and marketing, all over 35 y.o., all married with children.

So why all of us, respected professionals, with family obligations, leave behind the tranquility of “normal” life and jump into the Arena?

My personal answer to this, is that at certain point a man who always wanted to change the world understands he can. It is a happy coincidence it happened to us individually at the same time, so that we can do something about it collectively.

P.S. And in Israel, where it doesn’t snow and rails not so much, underground garage is a rarity.

Question on GData forums about Silverlight

That was the question asked on Feb 6, 2008 on GData forums.
I have blogged about it in my Silverlight Blog (that is being deprecated, as I aggregate my disparate blogs back into one)

Here is my post, relevant to about 2 month ago.

I am using Silverlight Alpha 1.1 with Google Spreadsheets. You can check it out at
www.gganttic.com
It is a Gantt chart based project management that stores tasks in Google Spreadsheets. The Gantt control is SL 1.1 based.

SL currently does not work cross domain. I hit this limitation several months back.
So I started using GData .NET client library on the server end to proxy.
it was said (don’t remember where) that SL 1.1 when released (it is going to be released as SL 2.0) will do cross-domain.
I did not want to wait for it. So here is the better way.

It is to do the communication in JavaScript layer, above SL.
SL is good in communicating between control and outside DOM environment.
So SL can throw an event, that is wired to a GData function, this function calls asynchronously a server (which can be any domain)
receives JSON data, calls into SL control via SL scriptable methods.
Scriptable are methods of SL that are marked so (if you are C#, it is [Scriptable] attribute on top of a method)

There is one drawback to it, the parsing logic would have to be in JS.
So for the time being I am doing it still with my server. There is a lot of guesswork involved in trying to understand which column of spreadsheet means what metadata of a task – so it is best done in C#.

A word of warning. Just like Frank (the guy on GData team) suspected, SL has a tiny subset of .NET framework. So GData .NET library would not even compile – it would need core classes SL has not gotten. It makes sense – SL is only 4.5 M download so Microsoft had to cut on everything.
I initially hoped I could use Netika control. Netika guys did a great work adding virtually all of Windows.Forms classes to SL.
It did not work. First Netika control is 900K, too much for my taste, and second it did not compile anyway. Incomplete implementation.

So the bottom line:
- In the future you’ll be able to call GData feeds direct from SL, but you’ll have to modify .NET library or write your own parser for the feeds.
- there is a good way of communicating cross domain using JavaScript library. GData has got JS lib for Calendar and Blogger, if I am not mistaken.

Silverlight 2.0 Beta 1 impressions

I am working to port our project management application to Silverlight 2.0 Beta 1.
I am almost done. Though I have not touched some areas of SL2, I touched enough to form an impression.

SL2 is a good successor to SL1.1, delivering on all the promises made.
It is in Beta, which means you can find bugs and rough edges without really trying that hard. But it is a stable product which I find I would use in production.
For all the problems I had, I found solutions very quickly.

I shall post my more detailed review in the future posts. I just this out quickly. If you develop Rich Internet Application, SL2 is ready. There may be other thing to consider when you make the choice, but technology is robust enough to be in the short list.

Google and Microsoft and how I came to write my Google-Microsoft hybrid application

The story of my online project management application (RIA or rich internet application) is this. I managed development team of a startup company.
I was (still am) a fun of spreadsheets way of managing tasks and schedules)

One day my CEO liked to visually see where we are, all in real time. I could have imported excel to MS Project, but that is not real time.
So the CEO hired a consulting company to set up Microsoft Project Server. That was done. We as a team struggled with all kinds of problems until about one month later we killed the idea, cut losses and I went back to Excel.

Yet something in an idea of collaboratively manage task list over the web clicked with me and I started to sketch the list of requirements.

Top one (#1) is – easy to use, instant to start using. It should be PM for the rest of us. Many people who manage projects are not professional project managers.
And people who are being managed, those people who actually make projects succeed or fail, should have zero resistance to the tool. I don’t ask for love. That will come later. I only ask for getting past all the defenses we put against anything new.

Then, it should be interactive. This means it is online, accessible from everywhere by everyone simultaneously if they want to. And it means that whatever you do you see the result at once. That is #2.

Integrated with a Calendar that other people can see is my #3. Projects are about people and time. Time one can dedicate to a task is really depends on what else that one is doing. Is she on vacation? Is she having 7 hours of unrelated meeting on every average day? Is there a holiday in India where you outsourced that particular task?

#4. Integrated with Communication tools and Documents system. Projects are about people and communication. People communicate in two ways – instantly and via capturing knowledge and thoughts in documents. So my ideal PM tool should work very naturally with email, phone, IM creating context for communication. And it should facilitate right documents being created and quickly accessed.

#5. It should be visual. Easy to grasp at one glance. I don’t know if this qualifies as a separate point. May be it is part of “instant” and “easy”.

These are my 5 points. Not many. But they are hard to do right. Many tried. All I have seen, and I am pro actively looking all the time, fail.

I think I can do better because I borrow from the excellent tools Google and Microsoft collectively provide.

How do you make something easy? You cut on all the fat you can find in terms of functionality. And you use the tools and metaphor the people are already used to.
Minimalistic approach is characteristic to Google Applications. I like that. And the applications are similar. I like that a lot. Every time I find differences between Documents and Spreadsheets I want to cry.

I see many posts in various blogs about the lists of RIA tools that small businesses can use. My biggest problem is that those tools each use their own UI conventions and design. So, I say to myself, my PM tool should blend into “Google Office”. It will remove the obstacle of learning UI for those who use it already. With Gmail and Calendar so popular, I get a free ride on UI familiarity front.

What I wish Google would improve is the speed. And that I doubt will happen. That is because Goolge is tied to AJAX, HTML and Javascript. It is not interactive enough. And if it is a concern for a text tool, it is a dead spell for a visual tool.

That is where Silverlight comes to rescue. The application, compiled, together with as complex logic as you care to program, resides on a client. There is a download time, but other then that ( and there are ways to help that too) it is every inch is as instant as your desktop application. Flash was another candidate, but lost in a match, and I’ll write about it in the future.

Another part of “instant” is to integrate with Google Spreadsheets (GS).
Many people use Excel (and I foresee many will use GS) for task management. That is because Microsoft Project is too complex. Excel on the other hand is instant. Simple. Easy. I was doing it forever, it now seems, yet felt guilty about it. Until I read Joel Spolsky’s post on how to use Excel for painless scheduling.

So my PM application message to the world would be – using Spreadsheets for PM is OK. The list a convenient metaphor. Timeline with milestones is just another metaphor. Both are fine. Sometimes you need one more then the other. It is good to have both. Spreadsheets gives you one. gGanttic gives you the other.

That is where I am today – developing a PM application in Silverlight on top of Google Applications (Spreadsheets for tasks now, and Calendar for time management couple of versions down)
I find Silverlight technology, C# 3.5 language and Visual Studio tools an excellent set to develop with. I find Google Apps an excellent platform to built on top of.

Some say Microsoft and Google are two poles of todays internet.
Yet I find it easy to benefit from whatever each does best. I am a bipolar developer.