English /
CMSSurveyReportMagnolia
Review report of Magnolia
Person(s) reviewing it: gremid
Official Information: Project Website - Features of Community Edition - Download - Documentation
Software version(s): current stable version is 2.1.5, reviewed version is upcoming 3.0 release, aka. 3.0-rc3
Status of the report: completed
What We Have
| Feature | Score (0-5) | Remarks |
|---|---|---|
| Anonymous open publishing | 1 | Magnolia is a traditonal enterprise web content management system. Thus it has no builtin functionality for open publishing workflows like those, indymedia sites are based on. Nevertheless such a workflow could be implemented on top of Magnolia's core, as is shown for example by the Java develtopment community portal InfoQ. It is based on Magnolia content-wise and has added community features like commenting. Anonymous article submission has not been implemented yet, but seems to be on their roadmap. |
| easy mirroring capability | 5 | Magnolia is based on a publish/subscribe-oriented architecture. This means, you can use one instance for content submission and editing, which in turn may publish committed content to several live instances, that deliver it simultaneously. |
| syndication -out/in | 1 | Not builtin. But might be added via one of several RSS frameworks available for Java, e.g. Jakarta Commons Feedparser. |
| search | 3 | The underlying content repository of Magnolia supports fulltext search over the whole content, including PDF-, MS-Word and -Excel-Documents and based on Apache Lucene. |
| The ability to create multiple instances | 3 | Magnolia suggests one installation per site, so multiple sites per instance are not supported. (Though there are ways to setup an instance serving multiple domains.) But as Magnolia is a standard JEE web application (like MIR), multiple instances can coexist in the same JEE container (for example Apache Tomcat) on the same machine. |
| Multimedia handling | 2 | On the one hand, any multimedia artifact can be uploaded and served as a binary object. Also image manipulation like resizing is supported with minimal programming effort. On the other hand, Magnolia does not offer specialised handling of multimedia assets (image gallery generation, ID3-tag parsing etc.) like other CMS |
| Categories | 1 | Not builtin, but implementable on top of Magnolia. See again InfoQ. |
| Good performence on affordable hardware | 1 | I suppose, Magnolia cannot compete with scripting-language-based CMS regarding the hardware cost/performance ratio when measured against low-volume sites. It needs up to 0,5 GB of RAM just to bootstrap in certain configurations. Anyway, given reasonable server hardware, it should scale quite well because of its content caching and its distributed architecture. For a reference site mentioning performance stats, see Amgen Tour of California. |
| Customisability | 5/2 | Regarding the site layout and the content structure, every aspect can be customized. I think, this is one of the real strengths of Magnolia: The ability to define the layout and the content structure quite flexibly, while retaining an easy-to-use, WYSIWYG-style content editing UI. Regarding the customization of the CMS itself, it depends on the amount of Java knowledge available. Magnolia uses JavaServer Pages directly for templating and the (relatively new) Java Content Repository API for defining content structure, UI dialogs and configuration settings. |
| internationalisation | 2 | Magnolia supports multi-language sites, though this is done via separate subsections and routing to the corresponding section depending on the user's choice. This might not be suitable for Indymedia sites out of the box, because of the need for a translation of distinct content fragments (aka. articles, news etc.). The best solution to this problem could be the definition of corresponding content types in Magnolia, supporting several variants of content, which are chosen based on the language setting of the site visitor. |
| translation | 5 | Magnolia has been translated in 15 languages. More translations can be added. |
| comments | 1 | See "Anonymous open publishing". |
| anti-abuse measures | 1 | See "Anonymous open publishing". |
| easy moderation | 1 | See "Anonymous open publishing". |
| calendar | 0 | Not available. Has to be implemented on top of the core code. |
| features | 4 | Feature-rich with regard to classic CMS functionality. |
| documentation | 2 | Magnolia has a vivid community but in part lacks necessary system documentation. It's easy getting started, but to customize the CMS, it's often: "Use the force, read the source!" |
What We Want
| Feature | Score (0-5) | Remarks |
|---|---|---|
| Logins | 3 | Magnolia features a user/role-based authentication system. Users and roles can be defined as needed. |
| access control | 3 | Authorization is based on users/roles and grant of access permissions to content resources, organized in a hierarchical tree. Whether this can fit Indymedia's need for access control has to be tested. |
| user moderation | 1 | Not builtin. Can be build on top of the given code, especially the given user base. |
| open editing | 1 | See "Anonymous open publishing". |
| profile | 2 | Basic user profile data are managed for CMS purposes, but they will have to be extended significantly to match typical requirements of a community-oriented portal. |
| user notifications | 2 | A scheduling system is built into Magnolia and the content repository supports notification of clients upon content modification. But as with all other community-oriented features, this has to be implemented and is not supported by Magnolia out of the box. |
| notify moderator button | 2 | See above. |
| podcasting/vodcasting | 0 | Not supported. |
| redundancy (DB content storage) | 5 | As already described above, redundancy can be achieved by publishing content to several live instances. Alternatively, the content repository of Magnolia supports several backends (file-based, BerkleyDB-based, Java embedded database engine, external RDBMS, custom repository implementation), so redundancy can also be provided by the underlying data store. |
| version control | 3 | The content repository supports versioned content, but as far as I know, Magnolia currently does not use this feature. This means there is no "undo" etc. |
| customisable skins by user | 2 | Not builtin, but should be solvable, because Magnolia supports different skins (templates). |
| accessability | 4 | The whole source delivered by Magnolia can be customized, so adherence to accessibility standards is not impeded by Magnolia. |
| xhtml validation | 3 | Magnolia generates XHTML-valid code out of the box. |
| GIS | 1 | Not supported by builtin functionality. |
| photo galleries | 1 | Not supported by builtin functionality. |
| licensing options | 4 | The community edition is licensed under the LGPL. |
| image manipulation | 2 | Basic functionality (resizing) is supported, but nothing sophisticated. |
| p2p integration | 0 | None. |
| social networking / filtering systems | 0 | None. |
| wysiwyg | 5 | Magnolia's WYSIWYG editor is based on FCKeditor. Every area of a page can be defined as being editable and can be modified by a simple point-and-click UI. |
| tagging: | 1 | Not supported out of the box, but implementable as shown on InfoQ. |
| anti-bot systems like captchas | 1 | See "Anonymous open publishing". |
| easier installation | 2 | Not as easy as e.g. a LAMP-based system, but not harder than MIR. |
| cross site search | 2 | Not implemented out of the box, but the underlying content repositories of Magnolia instances provide a public API, over which federated search could be implemented. |