Several people inquired as to how Quuve was built. Mariano wrote up a nice description of the various tools that were used to build Quuve -- see below. Note also that DebrisPublishing.com website is now up; we are still working on videos, but the site is very lean at present. -Cam Quuve Technologies Used Quuve has been and still is developed with Pharo. We try to use the most up to date stable version of Pharo. Quuve is deployed in GemStone. Package dependencies are managed via Metacello. Not only do we have many dependencies but also multiple Smalltalk environments: Pharo and GemStone. We also use Metacello to correctly load the needed packages and versions for each of the platforms. We use Seaside as our web framework and Twitter Bootstrap for Seaside for the layout and design of the app. While some parts are still using the full-request approach, most of the app is done with AJAX and only refreshing the needed part of the screen. We use JQuery a lot and some external JS libraries such a DataTables (for our table/reports), floatThead, HighCharts (for making charts), TinyMCE (to edit notes and a few other things), etc. We have a full-featured CRUD system based on Magritte. Actually, we have extended Magritte so much that all our descriptions and components are actually subclasses of the original Magritte classes. We have our own descriptions, components, report column classes, mementos, table layouts, etc. All our framework is Twitter Bootstrap integrated, fully AJAX and with lots of rich features. We can make a full CRUD report/form out of Magritte descriptions in no time. For persistency we have a library/layer which gives a unified way of saving and querying objects using a Glorp-like interface; it differs in that the where clauses are straight Smalltalk blocks. So far, when we work on Pharo we use Fuel as the "database", while for production we use GemStone. We have developed some compatibility packages so that same code would work in both Smalltalks in the way we expect. Quuve briefly used Oracleâs relational DB for persistence of its securities DB, connecting to it via the Glorp toolkit; because many mature financial DBâs are delivered in relational form, we may need to do so more extensively in the future. Although the need for this may subside as more data becomes available over REST-server interfaces. We use NeoCSV for parsing CSV, XMLParser, Zinc-REST and STON for providing REST web services, and SIXX for moving data between Pharo and GemStone. As to the web server, in Pharo we use Zinc, and in GemStone we use nginx + FastCGI. Quuve clients could be for example, an individual or a large advisory firm. In each case, Quuve is deployed as a "site/instance" for each customer. Each site is a fully isolated and secured instance of Quuve. Quuve is hosted in a cloud virtual hosting which allows us to scale as much as we need and we can have as many nodes as required. The operating system is CentOS Linux. We correctly perform backups, cleanings and maintenance of our sites. We have scripts to automatically create new sites very easily.